(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Patient doses and image quality in interventional neuroradiology"

PATIENT DOSES AND IMAGE QUALITY 
IN INTERVENTIONAL NEURORADIOLOGY 



By 
NIKOLAOS A. GKANATSIOS 





















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL 

OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT 

OF THE REQUIREMENTS FOR THE DEGREE OF 

DOCTOR OF PHILOSOPHY 

UNIVERSITY OF FLORIDA 

1998 



ACKNOWLEGEMENTS 

I would like to gratefully acknowledge the following who have helped me 
throughout my graduate work: 



Dr. Walter Huda, my advisor, for his invaluable guidance, time and patience 
throughout the course of this research and the preparation of this dissertation. 
I am very grateful for his continuous advice and suggestions throughout my 
graduate work. 

My Ph.D. committee members, Prof. James S. Tulenko, Dr. Wesley E. Bolch, 
Dr. Janice C. Honeyman, Dr Keith R. Peters and Dr. Irvine F. Hawkins, for 
reviewing my progress and guiding me through my Ph.D. research. 

Ms. Lynn Rill, for her valuable time evaluating all the radiographic images 
and for her review and comments on the manuscript. 

Mr. Dennis Pinner from Toshiba America Medical Systems, for his valuable 
insights into understanding the imaging equipment and for providing me with 
the requested information and documentation on the imaging system. 



• 



The Department of Radiology, for giving me the graduate assistantship to 
pursue my graduate studies, and for all the resources they made available for 
me throughout my graduate research. 

My beloved parents, Anastasia and Argyrios Gkanatsios, for their love, 
encouragement, and support throughout all my endeavors. They are the ones 
who made this possible for me. 



u 



TABLE OF CONTENTS 

ACKNOWLEGEMENTS ii 

LIST OF TABLES vi 

LIST OF FIGURES viii 

ABSTRACT xi 

CHAPTERS 

1 INTRODUCTION 1 

Interventional Neuroradiology 1 

Patient Dosimetry 2 

Dose Monitoring Systems 4 

Image Quality 5 

Purpose of This Work 6 

2 LITERATURE REVIEW 8 

Introduction 8 

Interventional Neuroradiologic Procedures 8 

Deterministic Radiation Effects 9 

Stochastic Radiation Effects 12 

Dosimetry 13 

Surface Dose 13 

Energy Imparted 15 

Effective Dose 18 

Image Quality 21 

Image Contrast 21 

Image Noise 23 

Spatial Resolution 26 

Imaging Technique Factors 27 

Tube Potential 28 

Input Exposure to the Image Receptor 29 

Magnification 30 

3 SURFACE DOSES 32 



in 






Neuroradiologic Imaging 32 

Clinical Practice 32 

Imaging Equipment 33 

Operation 35 

Imaging Techniques 36 

The Patient Dosimetry System 38 

System Description 38 

Calibration 40 

Evaluation 42 

Data Acquisition 45 

Fluoroscopy 47 

X-Ray Beam Localization 47 

Surface Doses 48 

Surface Dose Rates 49 

Fluoroscopic Times and Intervals 51 

Radiography 53 

X-Ray Beam Localization 53 

Surface Doses 54 

Surface Dose Rates 54 

Radiographic Frames 56 

Conclusions 57 

4 ENERGY IMPARTED AND EFFECTIVE DOSE IN NEURORADIOLOGY 60 

Introduction 60 

Method 62 

Energy Imparted 62 

Adult Effective Doses 66 

Pediatric Effective Dose 69 

Adult Patient Doses 71 

Energy Imparted 71 

Effective Doses 74 

Pediatric Patient Doses 75 

Energy Imparted 76 

Effective Doses 79 

Discussion 80 

Conclusions 83 

5 IMAGE QUALITY 86 

Image Acquisition 86 

Phantom Description 86 

Acquisition of Digitally Subtracted Images 88 

Dosimetry and Image Quality 94 

Dosimetry 94 

Image Quality Evaluation 95 

Precision of Measurements 96 

iv 



Results 97 

Tube Voltage 97 

Image Intensifier Input Exposure 100 

Geometric Object Magnification 102 

Discussion 107 

Patient Surface Dose 107 

Energy Imparted 110 

Image Quality 113 

Conclusions 114 

6 CONCLUSIONS 116 

Patient Dosimetry 116 

Surface Doses 1 16 

Effective Doses 117 

Image Quality 119 

Future Work 120 

BIBLIOGRAPHY 122 

BIOGRAPHICAL SCHETCH 131 



LIST OF TABLES 

Table page 

2-1. Deterministic Effects of the Skin after Single-Fraction Irradiation 11 

3-1 . List of the Input Signals Interfaced to the PEMNET Dosimetry System from 

the Toshiba Neurobiplane Imaging Unit 40 

3-2. Experimental Arrangements for Evaluating the Patient Dosimetry System 44 

3-3. Summary of the Ratios of the Measured to Calculated Surface Doses, XJX C , 

Obtained During Testing of the Accuracy of the Patient Exposure System 45 

4-1 . Computed a and /? Coefficients and Half- Value Layers for X-Ray Beams as 

a Function of Tube Voltage 64 

4-2. Backscatter Fractions of Radiation Exposure at Different Tube Voltages 65 

4-3. Patient Thickness and Area of Exposure Corresponding to the Head Region 

of Different Age Groups 66 

4-4. Patient Thickness and Area of Exposure Corresponding to the Trunk Region 

of Different Age Groups 67 

4-5. Values of Effective Dose per Unit Energy Imparted, Els in mJ/Sv, for 

Different Body Projections as a Function of Tube Voltage 70 

4-6. Standard Patient Mass for Different Age Groups 70 

5-1. Iodine Contrast Concentration in Each Vessel of the Vessel Insert 89 

5-2. Imaging Techniques During Tube Voltage Experiments 92 

5-3. Imaging Techniques During Geometric Object Magnification Experiments 93 

5-4. Score Describing the Visibility of Each Iodine Contrast Concentration 95 

5-5. Tube Voltage Dependency at 120 u.R/frame 98 

5-6. Tube Voltage Dependency at 440 uR/frame 98 

vi 



5-7. Image Intensifier Input Exposure Dependency 102 

5-8. Geometric Object Magnification Dependency at 120 uR/frame 104 

5-9. Geometric Object Magnification Dependency at 440 p.R/frame 104 

5-10. Comparison of the Effects of Tube Voltage, Input Exposure and Geometric 
Magnification on the Surface Dose for a Range of Changes in Threshold 
Iodine Concentration at 120 uR/frame 108 

5-1 1 . Comparison of the Effects of Tube Voltage, Input Exposure and Geometric 
Magnification on the Surface Dose for a Range of Changes in Threshold 
Iodine Concentration at 440 uR/frame 108 

5-12. Comparison of the Effects of Tube Voltage, Input Exposure and Geometric 
Magnification on the Energy Imparted for a Range of Changes in Threshold 
Iodine Concentration at Low Input Exposures Ill 

5-13. Comparison of the Effects of Tube Voltage, Input Exposure and Geometric 
Magnification on the Energy Imparted for a Range of Changes in Threshold 
Iodine Concentration at High Input Exposures 11 1 






vn 






LIST OF FIGURES 



Figure page 

1 - 1 . Unsubtracted image (left) where anatomical details are mixed with diagnostic 
information. Digitally subtracted angiogram (right) where anatomical 
information has been subtracted to allow easier visualization of vasculature 2 

2-1. Key parameters that affect patient dose and image quality in x-ray imaging 28 

3- 1 . Histogram of surface dose contribution at different x-ray tube voltages from 
frontal plane (black bars) and lateral plane (gray bars) fluoroscopy for an 
average interventional neuroradiologic procedure 37 

3-2. Histogram of surface dose contribution at different x-ray tube voltages from 
frontal plane (black bars) and lateral plane (gray bars) radiography for an 
average interventional neuroradiologic procedure 38 

3-3. Calibration setup of the frontal plane (left) and lateral plane (right) using an 

RSD RS-235 anthropomorphic head phantom 42 

3-4. Sample page from the PEMNET database showing all recorded information 

for the frontal imaging plane 46 

3-5. Histogram distribution of surface doses for 175 patients from frontal plane 

(black bars) and lateral plane (gray bars) fluoroscopy 49 

3-6. Histogram distribution of surface dose rates for 175 patients from frontal plane 

(black bars) and lateral plane (gray bars) fluoroscopy 50 

3-7. Histogram distribution of fluoroscopic times to 175 patients from frontal plane 

(black bars) and later plane (gray bars) fluoroscopy 52 

3-8. Histogram distribution of fluoroscopic intervals for 175 patients from frontal 

plane (black bars) and lateral plane (gray bars) fluoroscopy 53 

3-9. Histogram distribution of surface doses for 175 patients from frontal plane 

(black bars) and lateral plane (gray bars) radiography 55 

3-10. Histogram distribution of surface dose rates for 175 patients from frontal plane 

(black bars) and lateral plane (gray bars) radiography 56 

viii 



3-1 1 . Histogram distribution of the number of radiographic frames for 175 patients 

from frontal plane (black bars) and lateral plane (gray bars) radiography 57 

3-12. Histogram distribution of the total surface doses to 175 patients from frontal 
plane (black bars) and later plane (gray bars) fluoroscopy and radiography 
combined 59 

4-1 . Values of « as a function of water phantom thickness for tube voltages of 60 

kVp, 80 kVp and 100 kVp 63 

4-2. Effective dose as a function of patient mass for one joule of uniform whole 

body irradiation 71 

4-3 . Histogram distribution of energy imparted to patients from use of fluoroscopy 

during interventional neuroradiologic procedures 72 

4-4. Histogram distribution of energy imparted to patients from radiographic 

acquisitions during interventional neuroradiologic procedures 73 

4-5. Histogram distribution of the total energy imparted to patients undergoing 
diagnostic angiographic and therapeutic embolization neuroradiologic 
procedures 75 

4-6. Histogram distribution of the total effective dose to patients from biplane 

neuroradiologic examinations 76 

4-7. Energy imparted as a function of patient mass from fluoroscopy during 

interventional neuroradiologic procedures on pediatric patients. Line shows 

the linear fit between energy imparted and patient mass 77 

4-8. Energy imparted as a function of patient mass from radiographic acquisitions 
during interventional neuroradiologic procedures on pediatric patients. Line 
shows the linear fit between energy imparted and patient mass 78 

4-9. Energy imparted as a function of patient mass from interventional 

neuroradiologic procedures on pediatric patients. Line shows the linear fit 
between energy imparted and patient mass 79 

4-10. Effective dose as a function of patient mass from interventional 

neuroradiologic procedures on pediatric patients. Line shows the linear fit 
between effective dose and patient mass 80 

4-11. Comparison of Ele values vs. patient age as determined by Equation (4.4) and 

by using the dosimetry data from Hart et al. (1996a) 84 

5-1 . Schematic diagram of the acrylic phantom with the vessel and blank inserts 
used to simulate small vessels for the purpose of evaluating image quality in 
neuroradiology 87 

ix 



5-2. Experimental setup for DSA acquisitions 90 

5-3. Position of the two ionization chambers relative to the vessel insert (left). 

Subtracted image (right) 90 

5-4. Surface dose and energy imparted as a function of tube voltage 99 

5-5. Threshold iodine concentration as a function of tube voltage. The circles 

correspond to the 120 uR/frame and have been fitted to kVp 20 ' '. The squares 
correspond to the 440 pR/frame and have been fitted to kVp 156 100 

5-6. Surface dose and energy imparted as a function of image intensifier input 

exposure at 70 kVp 101 

5-7. Threshold iodine concentration as a function of image intensifier input 

exposure for a constant video level at 70 kVp 103 

5-8. Surface dose and energy imparted as a function of image intensifier input 

exposure at 70 kVp 105 

5-9. Threshold iodine concentration as a function of geometric object 

magnification at 70kVp 106 

5-10. Change in surface dose versus change in threshold iodine concentration 

with tube voltage input exposure and magnification 109 

5-11. Change in energy imparted versus change in threshold iodine concentration 

with tube voltage input exposure and magnification 112 



Abstract of Dissertation Presented to the Graduate School 

of the University of Florida in Partial Fulfillment of the 

Requirements for the Degree of Doctor of Philosophy 



PATIENT DOSES AND IMAGE QUALITY 
IN INTERVENTIONAL NEURORADIOLOGY 

By 

Nikolaos A. Gkanatsios 

December, 1998 

Chairman: James S. Tulenko 

Cochairman: Walter Huda 

Major Department: Nuclear and Radiological Engineering 

Diagnostic and therapeutic interventional neuroradiologic procedures involve 

imaging of catheter manipulation and vascular anomalies of the brain and generally 

require extensive use of x-ray radiation. Knowledge of the surface dose allows one to 

estimate the probability of inducing deterministic effects, whereas the corresponding 

value of effective dose is related to the patient stochastic risk. Modification of key 

imaging parameters (i.e., tube voltage, input exposure to the image receptor and 

geometric magnification) impact on patient doses and image quality, with the latter being 

defined as the lowest concentration of iodine in a vessel that may be visually detected in 

the radiographic image. A dosimetry system was installed on a biplane neuroradiologic 

imaging system to determine the doses to patients undergoing interventional 

neuroradiologic procedures. The dosimetry system computed surface doses on the basis 

of selected technique factors and information about patient location relative to the x-ray 



xi 



tube. The energy imparted to the patient, s, was determined using the surface dose, x-ray 
beam quality (i.e., kVp and HVL), exposure area and thickness of the patient and was 
converted into the corresponding value of effective dose, E. Values of surface dose and E 
were obtained for 175 patients, consisting of 149 adults and 26 pediatrics. Median values 
of surface doses to the head region were 1 .2 Gy in the frontal plane and 0.62 Gy in the 
lateral plane. Median values of the effective doses were 36 mSv for adult patients and 44 
mSv for pediatric patients. An acrylic phantom with 1-mm diameter vessels filled with 
iodine contrast was used to evaluate the effects of varying imaging parameters on signal 
detection and patient doses during digital subtraction angiography. Reducing the x-ray 
tube voltage offered the largest improvement in image quality for a given increase in 
patient dose. Increasing the image intensifier input exposure beyond 250 p.R/frame 
provided very little improvement in image quality, and this II exposure level should not 
be exceeded in interventional neuroradiologic imaging. A linear relationship was 
observed between magnification and threshold concentration, which offers significant 
patient benefits when surface doses are not expected to exceed the threshold doses for the 
induction of deterministic effects. 





Xll 











CHAPTER 1 
INTRODUCTION 



Interventional Neuroradiology 

Neuroradiology is a multi-imaging science, which utilizes all imaging modalities 
(i.e., plain film, digital radiography, computed tomography, magnetic resonance imaging, 
nuclear medicine, etc) to accomplish a complete diagnosis of human neurology. 
Neuroradiology can be distinguished as conventional or interventional neuroradiology. 
Conventional neuroradiology uses modalities such as plain film radiography, computed 
tomography (CT), magnetic resonance imaging (MRI) and ultrasound (US) to diagnose 
neurologic abnormalities. Interventional neuroradiology studies the vasculature and 
blood kinetics of the brain by means of catheterization performed with the transfemoral 
artery technique. Interventional neuroradiologic procedures can be further distinguished 
as diagnostic angiographic or therapeutic embolization procedures. The imaging portion 
of any interventional neuroradiologic procedure is accomplished by use of digital 
subtraction angiography (DSA). In digital subtraction angiography, a mask image is 
being subtracted from an image enhanced with injected iodinated contrast to isolate 
vasculature structures from the rest of the anatomy as shown in Figure 1-1. 

Interventional neuroradiologic procedures often involve long fluoroscopic 
exposure times and the acquisition of a large number of radiographic images. As a result, 
there is a possibility of induction of deterministic radiation effects such as skin erythema 

and epilation. It is also important to determine the stochastic risks involved in such 

1 



procedures in both adult and pediatric patients. Modification of key imaging parameters 
(i.e., tube voltage, input exposure to the image receptor and geometric magnification) 
impact on image quality and patient doses from interventional neuroradiologic 
procedures. The effects of these parameters on image quality and patient doses should be 
quantified and optimized in order to ensure adequate diagnostic image quality and 
reduced patient doses. 





FIGURE 1-1: Unsubtracted image (left) where anatomical details are mixed with 
diagnostic information. Digitally subtracted angiogram (right) where 
anatomical information has been subtracted to allow easier visualization of 
vasculature. 



Patient Dosimetry 
The surface dose is the dosimetric quantity that measures the dose absorbed in the 
surface of an irradiated region from radiation exposures. The surface dose accounts for 
the energy absorbed in the skin and can predict the possibility of inducing deterministic 
injuries from high dose interventional radiologic procedures (i.e., cardiac catheterization, 
abdominal interventional or neurointerventional procedures). Deterministic injuries 



associated with interventional neuroradiologic procedures primarily consist of injuries 
induced to the skin of the patient such as skin erythemas and epilations. Knowledge of 
surface doses may also provide information on the probability of deterministic injuries to 
the lens of the eye from interventional neuroradiologic procedures. 

The effective dose, E, is a dosimetric parameter, which takes into account the 
doses received by all irradiated radiosensitive organs. The effective dose is able to 
account for nonuniform irradiation of different organs and tissues in the body and can be 
used as an indicator of the stochastic radiation risk associated with radiologic x-ray 
examinations. Determining effective doses for radiologic examinations by measurement 
or calculation is generally very difficult. By contrast, the energy imparted, s, to the 
patient may be obtained from the x-ray exposure-area product incident on the patient. As 
energy imparted is approximately proportional to the effective dose for any given x-ray 
radiographic view, the availability of Els ratios (Huda and Gkanatsios, 1997) for common 
radiographic projections provides a convenient way for estimating effective doses. Such 
ratios primarily depend on the projection employed (body region irradiated and x-ray 
beam orientation) and secondarily on the tube potential and beam filtration. 

The effective dose as a dose descriptor in diagnostic radiology enables a direct 
comparison of the detriment associated with different radiologic procedures. Expressing 
patient doses in terms of the effective dose provides a consistent method of reporting 
doses from diagnostic radiologic examinations. Effective doses in interventional 
neuroradiology can simply be compared to other radiologic doses (i.e., computed 
tomography, nuclear medicine, cardiac catheterization procedures, etc.). The use of the 
effective dose also permits an estimate of patient risk to be obtained by using current 



stochastic risk factors (ICRP, 1991; UNSCEAR, 1993; NAS, 1990). Use of such 
stochastic risk factors with the effective doses computed for interventional 
neuroradiologic procedures will provide useful information on the stochastic risks to 
patients undergoing such high dose procedures. 



Dose Monitoring Systems 

A radiation monitoring system which provides feedback of dosimetric information 
could play a role in ensuring that patient exposures are as low as reasonably achievable 
(ALARA, ICRP, 1982). The benefits of a radiation monitoring system include 
identification of individual patients who may be at risk for the induction of deterministic 
radiation effects (Wagner et ah, 1994), provision of a formal record of the patient 
exposure as well as an increase in the radiologist's awareness of potential high patient 
doses. In addition, the radiation monitoring system can serve as a powerful tool to 
empirically investigate the tradeoffs between patient dose and corresponding image 
quality when radiographing appropriate phantoms. 

Use of modern on-line dosimetry systems on today's advanced x-ray imaging 
equipment provides the necessary tools for fast and accurate acquisition of dosimetry data 
on patients undergoing complex radiologic procedures. A patient dosimetry system 
(PEMNET*) was installed in 1995 in the neuroradiology suite at the Department of 
Radiology at Shands Hospital of the University of Florida. The patient dosimetry system 
monitored both frontal and lateral imaging planes and recorded the amount of radiation 



* Clinical Microsystems, Arlington, VA 



5 

received by patients undergoing interventional neuroradiologic procedures along with 
additional dosimetric information to help to compute effective doses. 



Image Quality 

The purpose of any radiographic image (analog or digital) is to provide the observer 
with adequate diagnostic information to detect and identify or rule out an abnormality and 
then to interpret its meaning and determine its cause. The ability of a radiographic image 
to convey this information to the observer depends on the quality of the image, which can 
be described in terms of contrast, noise and resolution. Image quality is very critical in 
interventional neuroradiologic procedures. The ability to visualize small and low contrast 
objects is of paramount importance, where neurovascular instruments may be as small as 
200 urn and where vessel sizes are as small as 100 urn. The produced images require 
high contrast, low noise and high resolution, which can be achieved with high radiation 
doses. Any dose reduction strategy must always ensure that image quality is not 
compromised and patients do not suffer any adverse clinical consequences as a result of 
inadequate visualization of catheters or vasculature. 

Ways to improve detection of small vessels during interventional neuroradiologic 
procedures using digital subtraction angiography (DSA) include the variation of major 
imaging parameters such as tube voltage and image intensifier input exposure, as well as 
use of geometric object magnification. Although these parameters affect image quality, 
they also influence patient surface doses and effective doses. Further study is necessary 
to improve our understanding of how technique parameters affect patient doses and to 
what extent they can improve image quality. 






Purpose of This Work 

Following the installation of the patient dosimetry system on the interventional 
neurobiplane suite at Shands Hospital of the University of Florida, dosimetry data on 
patients undergoing interventional neuroradiologic procedures were stored in a patient 
database for later analysis and evaluation. Dosimetric information on 149 adult patients 
and 26 pediatric patients who underwent interventional neuroradiologic procedures was 
recorded in the database. Seventeen of 149 adult patients and ten of the 26 pediatric 
patients recorded in the database underwent therapeutic embolization procedures. 

In this work, the dosimetry data to the adult and pediatric patients recorded by the 
patient dosimetry system are analyzed to compute surface doses to the patients' head 
region from interventional neuroradiologic procedures. Surface doses are then considered 
to assess the risk of deterministic effects to patients who undergo such interventional 
procedures, as well as similar high dose radiologic procedures. 

Information on the x-ray beam qualities (kVp and HVL) recorded by the patient 
dosimetry system with patient thickness and the x-ray beam exposure area are used to 
compute the energy imparted to these patients from recorded values of entrance skin 
exposures. Values of energy imparted are then converted to patient effective dose, E, 
using Els conversion factor corresponding to the projections and body regions irradiated 
during interventional neuroradiologic procedures. Values of Els for the posterio-anterior 
(PA) projections of the abdomen, chest and cervical spine and for the PA and lateral 
(LAT) views of the head are obtained from radiation dosimetry data computed using 
Monte Carlo calculations on an adult anthropomorphic phantom (Hart et ah, 1994a). 



This method is extended to determine effective doses to pediatric patients who differ in 
mass from the adult sized phantoms used in current patient dose assessment procedures. 

Manipulation of the tube voltage, input exposure to the image receptor and 
geometric object magnification impact on patient doses and image quality, with the latter 
being defined as the lowest concentration of iodine in a vessel that may be visually 
detected in the radiographic image. The effects of these imaging parameters on signal 
detection and the corresponding changes in patient doses are investigated in this work. 

The results of this work provide the radiologic community with a variety of 
information on patient surface doses, energy imparted and effective doses. Such 
information will help to evaluate the risks of deterministic and stochastic effects to 
patients undergoing interventional neuroradiologic or similar high dose radiologic 
procedures. The results on how imaging parameters (i.e., tube voltage, image intensifier 
input exposure and geometric object magnification) affect image quality will help to 
improve image quality and reduce patient doses, thus providing improved patient care to 
the healthcare community. 



CHAPTER 2 
LITERATURE REVIEW 



Introduction 



Interventional Neuroradiologic Procedures 

During diagnostic neuroradiologic procedures, all initial angiograms performed 
on a given vessel territory constitute complete coverage of arterial, capillary and venous 
phases. Subsequent examinations of that vessel with various alterations in positioning 
(projection), magnification, and contrast injection are performed to specifically evaluate 
the visualized or anticipated pathology. As a result, these are limited to arterial phase for 
aneurysms, capillary phase for tumors, and venous phase for study of venous patency. In 
therapeutic neuroradiologic procedures, a complete diagnostic angiographic procedure is 
followed by the introduction of embolic agents into the vasculature from a location next 
to the vascular abnormality. Such embolic agents might consist of gelatin sponge or 
polyvinyl alcohol for short-term occlusions or detachable balloons, metallic coils and 
cyanoacrylates for long-term occlusions. Subsequent evaluation of the pathology during 
embolization continues until a satisfactory occlusion of the vascular abnormality has been 
achieved. 

During all neuroradiologic procedures, frontal fluoroscopy is used in the truncal 
and thoracic regions to visualize catheter placement. Biplane fluoroscopy of the head 

8 



region is used for target position verification. Most DSA image acquisitions are 
performed using biplane acquisitions with only occasional use of single plane 
acquisitions. Single (frontal) plane imaging is primarily used to evaluate aneurysm neck 
origin with either standard DSA imaging or with rotational digital angiography (DA). 
Due to the nature of neuroradiologic procedures, good image quality, long fluoroscopic 
times and a significant number of angiographic images are required to evaluate any 
visualized pathology. Thus, neuroradiologic procedures result in high patient doses, 
primarily absorbed over the head region of the patient. This suggests the possibility of 
induction of deterministic radiation effects such as skin erythema and epilation (Huda and 
Peters, 1994; Shope, 1996). 



Deterministic Radiation Effects 

Deterministic or non-stochastic effects of ionizing radiation include the types of 
injuries resulting from whole-body or local exposures to radiation that cause sufficient 
cell damage or cell killing to substantial numbers or proportions of cells to impair 
function in the irradiated tissues or organs (ICRP, 1977). Since a given number or 
proportion of cells must be affected, there is a threshold dose below which the number or 
proportion of cells affected is insufficient for the defined deterministic injury to occur 
(ICRP, 1984). The threshold dose depends on the level of injury or the sensitivity of the 
tissues or organs being irradiated (Field and Upton, 1985). Any increase in dose above 
the threshold increases the level of injury, since fewer cells will survive at increased 
radiation dose. The effect will also increase with increased dose rate. Increased dose rate 



10 

will accelerate cell damage without allowing enough time for more effective cell repair or 
repopulation (ICRP, 1991). 

The doses that result in the clinical appearance of deterministic effects are 
generally of the order of a few Gray to tens of Gray. The time at which the effect 
becomes noticeable may range from a few hours to some years after exposure, depending 
on the type of effect and the characteristics of the irradiated tissue. The levels of 
radiation exposure and the irradiated tissues involved in interventional neuroradiology 
raise concern for deterministic effects of the skin and eyes. Table 2-1 lists the skin 
effects, threshold doses and time of observation of the expected effect after irradiation 
(Wagner et al, 1994). An early transient erythema may be observed a few hours after 
irradiation at skin absorbed doses in excess of 2 Gy, with a main erythema appearing 
about 10 days later, when skin doses exceed 6 Gy. A temporary epilation may be 
observed three weeks after an absorbed dose of 3 Gy to the skin surface with a permanent 
condition resulting at doses above 7 Gy. The concern to the eye involves small opacities 
on the lens of the eye, which may occur at doses of the order of about 1-2 Gy (Merriam 
and Focht, 1957; NAS, 1990). More severe cases of cataracts occur at thresholds above 
5-6 Gy with a latent period of about a year after irradiation. 

Deterministic effects will often have a more severe impact on children, since 
tissues are actively growing in comparison to adults (UNSCEAR, 1993). Additional 
deterministic effects that have been observed from irradiation during childhood include 
effects on growth and development, hormonal deficiencies, organ dysfunctions and 
effects on intellectual and cognitive functions. From current data available (UNSCEAR, 
1993), there is no evidence that the threshold of deterministic effects to the skin and eyes 



11 

are any different for children or adults. Although the brain is most sensitive to radiation 
insults during the first four years after birth when rapid growth and development of the 
brain takes place, single doses in excess of 10 Gy have to be administered to the brain 
during childhood before any deterministic effect of neurophysiologic or neuroendocrine 
nature are observed. 



Table 2-1: Deterministic Effects of the Skin after Single-Fraction Irradiation 



Deterministic 
Effect 


Dose Threshold 

(Gy) 


Onset Time 


Peak Time 


Early transient 
erythema 


2 


hours 


~24 hours 


Temporary epilation 


3 


~3 weeks 


— 


Main erythema 


6 


~10 days 


~2 weeks 


Permanent epilation 


7 


~3 weeks 


~ 


Dry desquamation 


10 


~4 weeks 


~5 weeks 


Invasive fibrosis 


10 


— 


~ 


Dermal atrophy 


11 


>14 weeks 


~ 


Telangiectasia 


12 


>52 weeks 


— 


Moist desquamation 


15 


~4 weeks 


~5 weeks 


Late erythema 


15 


-6-10 weeks 


— 


Dermal necrosis 


18 


>10 weeks 


~ 


Secondary 
ulceration 


20 


>6 weeks 


— 


SOURCE: 


Wagner et al, 1994. 







12 

Stochastic Radiation Effects 

Unlike the deterministic effects, stochastic effects are those for which the 
probability of an effect occurring is a function of dose without threshold and its severity 
of the effect is dose-independent (ICRP, 1977). Stochastic effects can be categorized as 
somatic (carcinogenic) effects and hereditary (genetic) effects, which may occur from 
injury to one or a small number of cells. Since a single cell may be enough to initiate the 
effect, there is a finite probability that the effect will occur however small the dose. 
Thus, stochastic effects are normally assumed to have no dose threshold below which the 
effect cannot possibly occur. 

Since a stochastic effect may occur at any level of radiation exposure, the 
exposure should be kept as low as reasonably achievable (ICRP, 1977). Unnecessary 
exposures should be avoided, necessary exposures should be optimized to provide the 
maximum benefit to the patient, and the total doses should be limited to the minimum 
amount consistent with the medical benefit to the individual patient (ICRP, 1982, 1983). 
In the case of optimizing medical procedures for the best dose-benefit outcome, the main 
concern should be the amount and type of information derived from the examination and 
its diagnostic value. 

Whole body irradiation or its equivalent as expressed by the effective dose 
equivalent or effective dose can be converted to a stochastic risk estimate using a total 
risk factor as determined by the ICRP (1977, 1978, 1991). From the ICRP (1991) 
attempt to estimate absolute stochastic risks from whole-body irradiation, a risk 
coefficient of 5x1 0~ 5 cancers and genetic abnormalities per mSv of radiation dose was 
derived. Such a risk coefficient puts one out of 20,000 people who received a whole 



13 

body dose equivalent of 1 mSv to risk of developing a fatal cancer. This is a more 
conservative value from the previously derived risk coefficient of 1.65xl0" 5 (ICRP, 
1978), where one out of 60,600 people who receive 1 mSv will develop a fatal cancer. In 
general, these risk factors need to be treated with great caution given the current 
uncertainties associated with the extrapolation of radiation risks from high doses to those 
normally encountered in diagnostic radiology (Fry, 1996; Puskin and Nelson, 1996) 

Although knowledge of the pediatric effective dose associated with radiologic 
procedure is helpful, it is important to note that any resultant detriment will depend on the 
age of the exposed individual. The stochastic radiation risks of carcinogenesis and 
genetic effects are generally greater for children than for adults to at least a factor of two 
(ICPR, 1991; NCRP, 1985). These factors would need to be taken into account when 
converting any pediatric effective doses into a value of risk or detriment. As a result, 
direct comparisons of pediatric doses with those of adults need to be treated with 
circumspection. 



Dosimetry 



Surface Dose 

The surface dose is the simplest and most frequent method used to measure 
patient doses from radiologic examinations because direct measurements on patients can 
be performed easily at the skin surface. The surface dose can be obtained from 
measurements of the skin exposure using an ionization chamber or specialized detectors 
attached to the skin surface during the examination (i.e., thermoluminescent chips, 



14 

fiberoptic scintillarors). The surface dose may also be converted to organ doses (Jones 
and Wall, 1985), although such an approach may result in errors of more than 20% 
(Padovaniefa/., 1987). 

Although simple to obtain, the surface dose is a poor indicator of the true 
significance of radiation exposure to the patient because it overlooks a number of 
important factors. For example, in a fluoroscopic exam the surface dose does not account 
for changes in the depth of the radiosensitive organs, changes in the exposed field size, 
changes in the position of the patient, and changes in the beam qualities, overlaying 
exposure fields and partial exposure of organs (Wagner, 1991). More importantly, the 
surface dose does not account for the area of exposure or the penetrating ability of the x- 
ray beam as the energy of x-rays varies. 

The above factors make surface dose a quantity of limited dosimetric value when 
estimating stochastic risks. However, the surface dose is the quantity of choice when 
trying to predict the occurrence of deterministic radiation effects of the skin during high 
dose interventional radiologic procedures. In this case, the surface dose is the dose to the 
organ, the skin. Vano et al. (1998) measured surface doses of 11-15 Gy resulting in 
erythematous lesions and chronic radiodermatitis from procedures in interventional 
cardiology. Huda and Peters (1994) computed an upper estimate of 6.6 Gy to the 
occipital region of the skull resulting in temporary epilation from an embolization 
neuroradiologic procedure. Other studies reported a range of surface doses for 
neurointerventional procedures. Norbash et al. (1996) studied twelve typical 
interventional neuroradiologic procedures and measured a range of 0.31-2.7 Gy to the 
skin surface of the head with a mean value of 1.5 Gy. Bergeron et al. (1994) measured 



15 

0.13-1.3 Gy with a mean value of 0.62 Gy for eight patients undergoing 
neurointerventional procedures. Chopp et al. (1980) reported an average surface dose to 
the head of 0.16 Gy. Gkanatsios et al. (1997) measured surface doses to 114 patients 
undergoing neurointerventional procedures and recorded doses ranging from 0.1-5.0 Gy 
with median values of 1 .2 Gy and 0.64 Gy for the frontal and lateral planes, respectively. 
Although about 30 patients in the latter study exceeded the deterministic threshold of 2.0 
Gy, no radiation-induced skin effects were noticed. 



Energy Imparted 

Although the surface dose or exposure has been popular when expressing patient 
radiation doses, these parameter do not take into account the x-ray beam quality (i.e., 
half-value layer) or the size of the irradiated area. An alternative quantity that can be 
used to assess patient dosimetry is the energy imparted, or integral dose (Wall et al, 
1979; Harrison, 1983; Huda 1984; Cameron 1992). Energy imparted is a measure of the 
total energy deposited in a volume (i.e., head, chest, abdomen, etc.) from exposure to x 
rays. The primary factors that affect energy imparted are the x-ray exposure, the area 
exposed, the energy of the x-ray beam, and the thickness of the exposed volume 
(Gkanatsios, 1995; Gkanatsios and Huda, 1997). Secondary factors affecting energy 
imparted are the filtration, the voltage waveform ripple, and the target angle (Shrimpton 
et al., 1984; Gkanatsios, 1995). Energy imparted may be used to compute the associated 
risk from different types of radiologic examinations, optimize imaging techniques with 
respect to patient dose, or even estimate the effective dose to the patient. 



16 

The computation of energy imparted can be carried out with accuracy and ease 
(Carlsson, 1963; Huda, 1984; Shrimpton et al, 1984; Gkanatsios and Huda, 1997). A 
number of approaches have been developed to obtain values of energy imparted from 
radiologic procedures (i.e., Carlsson, 1963, 1965a, 1965b; Carlsson et al, 1984; Harrison, 
1983; Shrimpton and Wall; 1982). Most methods calculate values of energy imparted 
from depth dose data or from estimates of the incident energy to the irradiated volume. 
Energy imparted generally depends on the x-ray beam quality, as well as the field size 
and irradiation geometry, which makes depth dose data of limited value in the everyday 
clinical setting. Monte Carlo techniques are another way to compute energy imparted, 
given that photon interaction cross sections and x-ray energy distributions are well known 
(Persliden and Carlsson, 1984; Boone 1992). However, these methods are computer 
intensive, time-consuming, and relatively cumbersome to use. Other simplified methods, 
such as the use of half-value thickness of tissue (Hummel et al, 1985), also can be used 
to calculate values of energy imparted. The most practical approach developed to obtain 
values of energy imparted is the use of transmission ionization chambers, which can 
generate energy imparted data from an exposure-area, or air collision kerma-area product 
(Shrimpton et al, 1984). Measurements of exposure-area product have been reported to 
result in an accuracy of energy imparted between 10% and 20% (Shrimpton et al, 1984; 
Berthelsen and Caderbland, 1991). However, exposure-area product meters do not take 
into account patient thickness, and the incident beam may not totally irradiate the patient. 
Although it may be possible to overcome both these limitations, an accurate and practical 
method for estimating energy imparted to patients that does not rely on special 
instrumentation would clearly be advantageous. 



17 

Recently, Gkanatsios and Huda developed a simplified method to compute energy 
imparted from any radiologic procedure (Gkanatsios, 1995; Gkanatsios and Huda, 1997), 
which may be used with the dosimetry equipment available in most radiology 
departments. The method is based on Monte Carlo calculations of energy imparted from 
monoenergetic photons (Boone, 1992) and makes use of published diagnostic energy x- 
ray spectra (Tucker et ai, 1991). The patient is modeled as a homogenous slab of water 
with a specified thickness. The water equivalence of a given patient may be obtained by 
direct measurement of the patient or by estimating the thickness of water which results in 
the same x-ray technique factors when the imaging equipment is in automatic exposure 
control (AEC) mode. Experimental measurements needed for this computation include 
the entrance skin exposure, the x-ray beam qualities (kVp and HVL), as well as the 
exposed area and thickness of the patient, all of which may be readily measured or 
otherwise estimated. Gkanatsios and Huda compared this method with values of energy 
imparted determined using Monte Carlo techniques and anthropomorphic phantoms for a 
range of diagnostic examinations. At 60, 80 and 120 kVp, absolute values of energy 
imparted obtained using this method differed by 3%, 10% and 22% respectively, from the 
corresponding results of Monte Carlo computations obtained for an anthropomorphic 
phantom. 

The assumption that energy imparted to the head and trunk can determine 
radiologic risk has been investigated by many researchers (Bengtsson et ai, 1978; Huda, 
1984; Carlsson and Carlsson, 1986; Le Heron, 1992; Chappie et al., 1994). It was found 
that there may be a valid relationship between energy imparted and radiologic risk. 
Although the radiosensitivities of different organs and tissues are ignored, the energy 



18 

imparted will predict associated radiologic risks as accurately as when computing doses 
to individual organs (Wall et ah, 1979; Harrison, 1983; Cameron, 1992). A reasonable 
linear correlation within a factor of two or three (IPSM, 1988; Huda and Bissessur, 1990) 
was also detected between total energy imparted and effective dose to the head and trunk. 
Provided that the examining view (AP, PA, LAT, etc.) and the x-ray beam qualities are 
known, the effective dose can be determined easily from values of energy imparted (Huda 
and Gkanatsios, 1997, 1998). 

Effective Dose 

The effective dose, E, is a dosimetric parameter, which takes into account the doses 
received by all irradiated radiosensitive organs. The effective dose is able to account for 
nonuniform irradiation of different organs and tissues in the body. Thus, the effective 
dose is considered a measure of the stochastic risk associated with radiologic 
examinations by directly comparing partial body irradiation to whole body radiation 
exposure (ICRP, 1977, 1991; Huda et al., 1991). Although the effective dose is an 
occupational dose quantity based on an age profile for radiation workers, this dose 
descriptor is being increasingly used to quantify the amount of radiation received by 
patients undergoing radiologic examinations which use ionizing radiation (ICRP, 1987; 
NCRP, 1989; UNSCEAR, 1993). 

Measurement or computation of effective doses for any x-ray examination is 
difficult and time consuming. An additional problem is that most measurements or 
calculations make use of a standard phantom based on the reference man as defined by 
the International Commission on Radiological Protection (ICRP, 1975). Although the 



19 

importance of patient size for medical radiation dosimetry has been recognized 
(Lindskoug, 1992; Chappel et al, 1995), it is not obvious how to scale the effective dose 
computed for standard man to different sized patients, such as pediatric patients, who 
undergo similar examinations. These limitations impede the wider use of effective dose 
in radiology. Huda and Gkanatsios (1997) proposed a method to determine the effective 
dose, E, to patients undergoing any radiologic examination using the energy imparted to 
the patient, e. Values of Els were obtained from the radiation dosimetry data presented 
for 68 x-ray projections computed using Monte Carlo calculations on an adult 
anthropomorphic phantom (Hart et al, 1994a). The energy imparted to patients may be 
determined from values of the exposure-area product incident on the patient and can be 
combined with Ele ratios (i.e., 5.0 mSv/J for a head PA view) to yield values of the 
patient effective dose. In addition, this method was extended to determine effective doses 
to patients who differ in mass from the adult sized phantoms used in current patient dose 
assessment procedures (Huda et al, 1989b; Le Heron, 1992). 

Although the computation of effective dose is cumbersome in most cases, a range 
of effective doses has been reported in the literature that pertain to neurointerventional 
procedures. Feygelman et al. (1992) studied ten cases and reported values ranging from 
1.6-14 mSv with a mean of 6.2 mSv. Bergeron et al. (1994) reported an average of 1.8 
mSv with a range of 0.44-3.4 mSv for a limited number of eight patients undergoing 
similar procedures. McParland (1998) reported a median of 7.0 mSv with a range of 2.1- 
20 mSv when he computed effective doses to patients undergoing cerebral angiography. 
A wider range was reported by Berthelson and Cederblad (1991), who computed effective 
doses between 3.5 mSv and 25 mSv. 



20 

Despite its popularity, the effective dose introduces some problems when used in 
diagnostic radiology. First, it does not account for differences between the age 
distribution of workers and that of the general public with regard to the determination of 
the appropriate organ weighting factors. The effective dose also excludes curable cancer 
or hereditary harm beyond the second generation. Both these factors make the effective 
dose a questionable quantity in risk assessment associated with diagnostic radiologic 
procedures (UNSCEAR, 1988; Cameron, 1992). It should also be mentioned that the 
effective dose applies only to low radiation doses, which generally is the case in 
diagnostic radiology. However, in areas like cardiology and neuroradiology, where 
extended diagnostic and therapeutic procedures may deliver local patient doses of several 
Gray, the effective dose may not be an appropriate dosimetric quantity. 

Another problem with the effective dose is the uncertainties involved with its 
calculation. The calculation of the effective dose must include an analysis of the dose 
distribution within the body, which is difficult to do for radiologic procedures, 
particularly fluoroscopy. As an alternative, dose distributions are derived from Monte 
Carlo techniques using mathematical phantoms (Gibbs et al, 1984; Jones and Wall, 
1985; Huda et al, 1991; Le Heron, 1992) or from calculations of the average organ dose 
in anthropomorphic phantoms (Faulkner and Harrison, 1988; Huda et al, 1989a, 1989b). 
Such techniques, though, can only provide approximations of the true organ dose 
distribution. Furthermore, the selection of the "remainder" organs is problematic in dose 
distribution analysis and may vary for each examination. Effective dose also requires the 
use of a dose equivalent, which is based on the quality factor, Q, of the type of radiation 



21 

involved (ICRP, 1977), and use of organ and tissue weighting factors, w^ (ICRP, 1991). 
Both these factors are considered to be biologically uncertain (Cameron, 1992). 

Notwithstanding the fact that there are problems associated with converting 
effective doses to a corresponding detriment (Huda and Bews, 1990), there are important 
benefits to be gained by using effective dose to quantify patient doses in diagnostic 
radiology. One advantage is that the effective dose attempts to measure the risk to the 
patient, which is the motivation for all patient dosimetry studies in diagnostic radiology. 
In addition, the effective dose to a patient undergoing any examination may be compared 
to that of any other radiologic procedure, as well as to natural background exposure and 
regulatory dose limits, which are increasingly expressed using effective dose values 
(ICRP, 1991; NRC, 1995a, 1995b). 



Image Quality 
The extraction of adequate diagnostic information from radiographic images is 
important in radiology in order to detect and identify an abnormality and then to interpret 
its meaning and determine its cause. Thus the quality of the radiographic image is very 
important in conveying diagnostic information to the observer. Image quality can be 
described in terms of contrast, noise and resolution. 



Image Contrast 

Image contrast can be defined as the difference in the optical density (film) or 
brightness (digital) in an image between an area of interest and its surrounding 
background. Image contrast is determined by several factors including the characteristics 



22 

of the materials being imaged, the characteristics of the x-ray spectrum, the 
characteristics of the detector and display media and physical perturbations such as 
scattered radiation (Hasagawa, 1991). These dependencies separate image contrast into 
radiographic (subject) contrast, detector contrast and display contrast. 

Radiographic contrast. Radiographic or subject contrast characterizes the 
differences in x-ray fluence emerging from different regions of the imaged object. 
Radiographic contrast depends on differences in material thickness, atomic numbers, 
physical density and electron density between different regions of the imaged object and 
their interaction with radiation. 

Detector contrast. Detector contrast on the other hand, can be expressed as the 
ability of the imaging detector to convert differences in x-ray fluence emerging from an 
object to differences in optical density (film detector) or brightness (digital detector). The 
detector contrast can shape the radiographic contrast according to the detector's 
characteristic response to x-rays. Thus, detector contrast depends on the properties of the 
detector material, its thickness, atomic numbers, electron density and the physical process 
by which the detector converts x-ray fluence into an image. 

Display contrast. The third component of image contrast is the display contrast, 
which refers to the digital display of images. Display contrast depends on the display 
parameters (i.e., window and level) under which the image is viewed and can be 
manipulated by the observer. 

Other contrast dependencies . Image contrast in general, is also affected by 
physical perturbations such as scattered radiation, image intensifier veiling glare and the 
base and fog of film, all of which reduce image contrast. 



23 

Image Noise 

Every radiographic image is degraded by noise superimposed on the image by 
random processes occurring along the imaging chain. Detection of a signal that is 
superimposed on noise depends on the relative magnitude of the noise compared to the 
signal and the ability of the observer to differentiate between the brightness distribution 
of the noise and that of the signal plus noise (Giger et al, 1986b). The overall noise of an 
image consists of various noise components. The statistical nature of x-ray production 
and attenuation in the detector results in quantum mottle. Structure mottle, electronic 
noise, quantization noise, time jitter and display device noise are additional noise 
components in digital imaging detectors (Giger, 1985). 

In digital imaging as in digital subtraction angiography, the noise components can 
be categorized as static and non-static noise. Static noise is independent from one frame 
to the next and always presents the same pattern. Thus, static noise is eliminated in 
digital subtraction angiography. Structure mottle is the most important static noise 
component in digital imaging. Non-static noise is frame dependent, which means that the 
noise pattern varies from one frame to the next. Non-static noise sources are always 
present in digital subtraction angiography. Significant no-static noise sources in digital 
imaging are the quantum mottle and electronic noise. 

The primary source of noise in digital imaging is usually quantum mottle, which 
corresponds to random spatial fluctuations of the distribution of x-ray quanta absorbed by 
the detector. Since the production and attenuation of x rays are Poisson statistical 
processes, quantum mottle follows Poisson statistics, which makes it easily quantifiable. 
Consequently, increasing the exposure to the imaging detector will improve a 



24 

radiographic image by decreasing quantum mottle. Improving the attenuation properties 
of the imaging detector will also reduce quantum mottle of radiographic images. 

Secondary sources of noise become important in radiographic imaging, when the 
image receptor is exposed to high enough radiation to eliminate most of the quantum 
mottle. Secondary noise sources in digital imaging consist of the structure mottle, 
electronic noise, quantization noise and time jitter. 

Structure mottle . Structure mottle is the second most important noise component 
in single- frame digital imaging after quantum mottle, and it becomes the dominant noise 
source in images acquired using high x-ray fluence (Giger et al, 1986b). The structure 
mottle is introduced to the imaging line by the image intensifier. Structure mottle 
depends on the physical structure of both the input and output phosphor layers. Since 
structure mottle is a static component of image noise, its noise pattern is constant from 
frame to frame. Therefore, structure mottle can be eliminated by the subtraction of two 
image frames as done in digital subtraction angiography. Another characteristic of 
structure mottle pertaining to its static nature is that structure mottle remains unchanged 
after frame integration. 

Electronic noise . Electronic noise arises from the video camera as a form of dark 
current added to the exposure-dependent video signal. The magnitude of electronic noise 
is inversely proportional to the dynamic range of the TV camera and is relatively 
independent of video signal size. In order to minimize the perturbations added to a 
digital radiographic image by electronic noise, the video signal should be maximized 
when possible (Cohen et al., 1982). In general, the electronic noise in a digital imaging 
system is quite small relative to the quantum and structure mottle (Roehrig et al, 1981; 



25 

Baiter et al, 1984). However, electronic noise becomes a significant noise source when 
an object is imaged at low video levels and using low x-ray fluence. It was also 
demonstrated by Geiger et al. (1986b) that electronic noise contribution becomes 
substantial at spatial frequencies of about 1.0 cycles/mm. 

Quantization noise . Another noise component of a digital imaging system is quantization 
noise. Quantization noise is the error introduced into an analog signal (i.e., TV video 
signal) when it is digitized. Quantization noise depends on the width of the quantization 
step. In general, digital imaging systems are designed to minimize quantization errors, 
which makes quantization noise insignificant in comparison to quantum mottle or even 
electronic noise (Burgess, 1984; Boon et al., 1990; Rajapakshe and Shalev, 1994; Baxter 
etal, 1997). 

Time jitter . Another component of noise that may appear in digital imaging systems is 
time jitter (Arnold and Scheibe, 1984; Esthappan et al., 1998). Time jitter is usually 
caused by incorrect alignment of the scanning electron beam in the television camera 
from one video frame to the next. Time jitter may also be caused by a variable 
asynchrony between the video signal and the analog-to-digital converter. In general, time 
jitter produces a variation in pixel position from one image frame to the next. The 
importance of time jitter becomes significant in digital subtraction angiography, when 
this spatial pixel shift changes the spatial pattern of static noise bringing up structure 
mottle in a digitally subtracted image. Therefore, careful design and stable electronics are 
required in digital imaging systems to avoid time jitter in order to eliminate structure 
mottle completely from digitally subtracted images. 



26 

Spatial Resolution 

The third parameter used to quantify image quality in addition to contrast and noise 
is spatial resolution, frequently referred to as resolution. Although spatial resolution does 
not have as much of an impact on image quality as contrast or noise, in applications of 
neurointerventional imaging spatial resolution becomes somewhat more important. 
During interventional neuroradiologic procedures, the need to visualize tiny 
neurovascular instruments (i.e., catheters and guide wires) and vessels as small as 100 
um, requires high spatial resolution. The spatial resolution of an imaging system can be 
characterized by its modulation transfer function (MTF) (Haus, 1979; Metz and Doi, 
1979) which can be obtained from measurements of the point or line spread functions. 
The determination of MTF of digital imaging systems, however, requires careful handling 
to avoid aliasing effects caused by the discrete data sampling of digital systems (Giger 
and Doi, 1984; Fujita et al, 1985). In general, the spatial resolution of an imaging 
system depends on geometric, motion, detector and digitization unsharpness. 

Geometric unsharpness . Geometric unsharpness refers to the loss of image detail 
due to the finite size of the radiation source (i.e., focal spot) (Hasagawa, 1991). Heat 
loading of the anode of an x-ray tube requires that the focal spot is large enough to 
dissipate the generated heat. The finite size of the focal spot creates unsharpness called 
penumbra at the edges of the imaged object. To limit the amount of geometric 
unsharpness in neurointerventional imaging, x-ray tubes with steep anode angles (i.e., 9- 
1 1 degrees) and small effective focal spots (i.e., 0.3 mm or 0.6 mm) are used. Another 
practice often used in neurointerventional imaging is the use of magnification, which also 
increases geometric unsharpness. 



27 

Motion unsharpness . Motion unsharpness refers to the loss of spatial resolution 
due to motion of the x-ray source, detector and/or object being imaged (Hasagawa, 1991). 
When one or more of these components move, motion unsharpness is introduced, which 
degrades spatial resolution. Patient motion caused by discomfort and the continuous 
moving of the patient's heart and diaphragm is usually the greatest concern, since source 
and detector can be easily secured in place. Sedation or immobilization of the patient 
during a radiographic procedure and short exposure times will help reduce the amount of 
motion unsharpness. 

Detector unsharpness . Detector unsharpness refers to the loss of spatial resolution 
due to the finite resolving power of the detector (Hasagawa, 1991). In screen- film 
systems, detector unsharpness is also caused by light diffusion in the intensifier screens. 
Thicker intensifier screens will allow more light diffusion and create more unsharpness. 
In digital imaging, spatial resolution depends on the TV bandwidth and pixel size. Thus, 
TV systems with 1024 lines are used in neurointerventional applications. In addition, any 
digitization will result in loss of spatial resolution due to the inherent pixellation of a 
digital image in comparison to the original analog image. 



Imaging Technique Factors 
Patient doses and image quality are both influenced by the selection of imaging 
techniques. Figure 2-2 shows some key parameters along the line of an x-ray imaging 
system which can alter patient absorbed doses and image quality. Such parameters are 
the tube voltage, tube filtration, input exposure to the imaging detector, magnification 
and image processing. With the exception of image processing, an attempt to decrease 



28 

patient dose by altering one or more of these parameters will also degrade image quality. 
Thus, tradeoffs between varying different imaging technique factors merit investigation 
to find better ways to improve image quality while maintaining low patient doses. 



&n> 





Image Processing 



IV "titty Input Exposure 
\C jI 

Filtration 




Figure 2-1 : Key parameters that affect patient dose and image quality in x-ray imaging. 



Tube Potential 

Very early in the history of diagnostic radiology, tube voltage and the use of 
specialized K-edge filters were studied extensively to optimize patient dose and image 
quality (Trout et al, 1952; Koedooder and Venema, 1985; Shrimpton et al, 1988; Nagel, 
1989). In general it was shown that an increase in tube voltage decreases patient 
exposure and degrades image quality. The optimal tube voltage for detecting large-area, 



29 

low-contrast iodinated objects was determined to be between 50-60 kVp (Tapiovaara and 
Sandborg, 1995). The same study also showed that for detecting thin, soft-tissue detail a 
tube voltage between 70-100 kVp should be used. Also, Thompson et al. (1983) 
concluded that high tube voltages between 100-110 kVp combined with increased 
contrast agent concentration are the optimal techniques for detecting stones in operative 
T-tube cholangiography. 

The optimal tube potential depends on the imaging requirements of each imaging 
procedure. In interventional neuroradiologic procedures, where both visibility of small 
iodinated vessels and high spatial resolution are important, low tube voltage may be used 
to maintain adequate image quality. As a consequence, low tube voltage will contribute 
to high patient absorbed doses. As the tube voltage increases, both entrance absorbed 
dose and energy imparted to the patient decrease for a constant input exposure to the 
imaging detector. However, it should be noted that for a constant input exposure to the 
patient, increase in tube voltage would increase the energy imparted to the patient 
(Gkanatsios and Huda, 1997). 



Input Exposure to the Imaee Receptor 

The relationship between input exposure to the image receptor and patient 
absorbed dose is linear. The input exposure to the image intensifier also affect image 
quality. As the input exposure increases, the dose to the patient increases and the 
significance of the quantum mottle in a radiographic image decreases. Since most 
radiographic images are quantum limited, increasing the exposure to the image receptor 
will always improve contrast-to-noise (CNR) and signal-to-noise (SNR) ratios by 












30 

reducing image noise. However, as the input exposure increases to the point that other 
noise sources (i.e., structure mottle in singe-frame digital radiographs) become as 
significant as quantum mottle, then any increase in input exposure will have a minor 
effect on image quality. 

Any increase in input exposure to the image receptor at a given tube voltage will 
increase patient absorbed doses, proportionally. For a film-screen imaging system, where 
the input exposure to the system is controlled by an optimum optical density, there is 
negligible flexibility in varying the input exposure. In digital imaging systems, however, 
the range of input exposure can vary considerably and still produce a useful, diagnostic 
image. Thus, while operating in the range of input exposures where quantum mottle is 
the dominant noise component, increasing the input exposure for the purpose of 
improving contrast visibility is justifiable. However, if the input exposure to the image 
receptor is already high enough so that quantum mottle is not the primary component of 
radiographic noise, any increase in input exposure only increases patient absorbed doses. 
Such practice lowers the standard of patient care by not following the ALARA principle. 

Magnification 

Magnification and its effects on image quality have been studied in both 
conventional radiography and mammography (Doi and Rossmann, 1974; Wagner et al, 
1981a, 1981b; Sandrik and Wagner, 1982). In general, magnification improves visibility 
of small, low contrast objects. As the magnification increases, the effective noise in the 
image detector is reduced improving the signal-to-noise ratio, and visibility of small 



31 

structures improves (Doi and Imhof, 1977). Scatter radiation is also reduced with 
increased magnification, which improved contrast detectability (Sandor and Nott, 1980). 

In neurointerventional radiologic procedures, magnification is often used as a tool 
to visualize small vasculature. Care should be taken, however, when magnification is 
used, since the entrance absorbed dose to the patient increases significantly with 
magnification. Energy imparted, on the other hand, is independent of magnification as 
both distance from the x-ray source and area of exposure decrease equally as 
magnification is employed. The choice between geometric — change of distance between 
patient and x-ray source — and electronic magnification -changing the input diameter of 
the image detector — should be considered every time magnification is required, and the 
possibility of dose savings between the two methods should be investigated in any 
imaging system. 



CHAPTER 3 
SURFACE DOSES 



Neuroradiologic Imaging 
Clinical Practice 

Interventional neuroradiologic procedures are performed on patients suspected to 
have vascular anomalies in the brain (i.e., aneurysm, vasculitis or arteriovenous 
malformations), patients that have brain tumors, patients who have had a stroke episode 
or patients requiring certain types of psychological evaluation. A neurointerventional 
procedure may be a diagnostic angiographic or therapeutic embolization procedure. In 
diagnostic angiographic procedures, the vasculature and blood dynamics of certain parts 
of the brain are studied by imaging the kinetics of radio-opaque media injected in the 
vasculature of the brain. In therapeutic neurointerventional procedures, corrective action 
is taken to occlude vascular anomalies by injecting embolic agents such as gelatin 
sponges or metallic coils. Usually, a therapeutic embolization procedure is preceded by a 
diagnostic angiographic procedure. In both types of neurointerventional procedures, x- 
ray imaging is used extensively in the forms of fluoroscopy, conventional film and digital 
radiography. 

The transfemoral artery technique is used to perform neurointerventional 
procedures, where a catheter is inserted into the common femoral or deep femoral artery 
from where it is driven to the vascular network of the brain. Limited amount of frontal 



32 



33 

plane fluoroscopy is used on the trunk and thoracic regions to guide the catheter up to the 
vertebral or carotid arteries. Once there, further use of fluoroscopy in both imaging 
planes, frontal and lateral, is used to position the catheter at the appropriate site to be 
studied. Although biplane fluoroscopy is used in this stage, most of the fluoroscopy is 
still done using the frontal plane. Once the catheter is in place, radio-opaque contrast is 
injected to that location and a series of radiographic images are acquired in plain film or 
in digital format. In diagnostic angiographic procedures, the acquisition of radiographic 
images is done in biplane mode almost exclusively. In therapeutic embolization 
procedures, both biplane and single plane imaging, either frontal or lateral are used 
during different stages of the embolization progress evaluation. During each radiographic 
acquisition, the frame rate and number of frames may vary from 1-3 frames per second 
and 10-50 frames per acquisition, respectively. 

Imaging Equipment 

The x-ray imaging system used in this study consisted of a biplane Toshiba t 
KXO-80 high voltage diagnostic x-ray generator and the Toshiba DFP-2000A/A3 digital 
fluorography system configured for neuroradiologic procedures. The configurations of 
the two imaging planes, frontal and lateral, were identical. The frontal plane was built 
around the Toshiba KXO-80C high frequency x-ray generator. The lateral plane was 
based on its sister generator, the KXO-80D. Both generators were interfaced together to 
function as a biplane unit suited for neurointerventional applications. 



+ Toshiba America Medical Systems, Tustin, CA 



34 

Tri-focal metal Toshiba ROT ANODE x-ray tubes having nominal focal spot sizes 
of 0.3 mm, 0.6 mm and 1.0 mm and inherent filtration of about 3.0 mm aluminum were 
used as the x-ray sources. The collimator assembly provided an almost circular x-ray 
field using a multi-blade collimating iris matched tightly to the size of the image 
intensifier input area. The collimator assembly provided total collimation with the help 
of four metal blades or partial collimation using wedge shaped, transparent filters. A 
support table with a comfort pad totaling an equivalent filtration of 3.0 mm aluminum at 
80 kVp were placed between the x-ray beam and the patient. 

Two image receptors were available. The first receptor was a biplane screen-film 
system rated as 600-speed and 400-speed for the frontal and lateral imaging planes, 
respectively. The second image receptor was a digital radiography detector. The digital 
radiography detector consisted of a Csl image intensifier tube with three effective input 
diameters of 31 cm, 23 cm and 15 cm. A carbon fiber interspaced grid with a ratio of 
10:1 was used to remove scatter radiation to the input phosphor of the image intensifier. 
An automatic iris control adjusted the amount of light reaching the TV camera. The TV 
camera consisted of a high-resolution CCD head (1024 lines) and 10-bit analog to digital 
converter. Digital information was passed from the TV camera to the digital image 
processor. Analog video signals of 1024 lines at 60 Hz interlaced were passed to the live 
fluoroscopic high-resolution monitors. The digital image processor was a Toshiba DFP- 
2000A/A3 digital fluorography system capable of split display fluoroscopy, roadmap 
fluoroscopy, digital angiography and digital subtraction angiography. 






35 

Operation 

The x-ray imaging system was capable of continuous fluoroscopy or pulsed 
fluoroscopy at 15 or 30 frames/sec. Pulsed fluoroscopy could operate at low or high kVp 
ranges when a high or low tube current (mA) was selected. Pulsed fluoroscopy at 30 
frames/sec and high mA setting was primarily used as the default fluoroscopic technique 
during most neurointerventional procedures. Targeted input exposures to the image 
intensifier in fluoroscopy were measured at 1.9 (j,R/frame, 3.4 jiR/frame and 4.7 
uR/frame for the 31 cm, 23 cm and 15 cm input diameters, respectively, using a 2.0 mm 
copper filter. 

Limited amount of frontal plane fluoroscopy was used on the trunk and thoracic 
regions to drive the catheter to the head region. On average, about thirty seconds (34 ± 
1 1 sec) of fluoroscopy were spent along the trunk region. An additional two minutes 
(133 ± 77 sec), on average, were spent along the upper thoracic, lower neck region to 
enter the vertebral or carotid arteries. The remaining use of fluoroscopy was allocated to 
the head region during placement of the catheter in the appropriate arterial branch to be 
imaged. During this time, the majority of fluoroscopy was performed in the frontal plane. 
Lateral fluoroscopy was used in those cases where frontal imaging does not contain 
adequate information to help in catheter manipulation. Biplane fluoroscopy was used to 
verify target positioning prior to each contrast injection and imaging. 

Digital subtraction angiography (DSA) was the primary imaging method during 
interventional neuroradiologic procedures and was mainly performed at a rate of 3.0 
frames/sec. Rates up to 6.0 frames/sec were used to evaluate high flow dynamics. The 
input exposure to the image intensifier in digital subtraction angiography was user 



36 

selected and it could vary from 50 uR/frame to 1000 uR/frame with 500-700 uR/frame 
being the default value. 

In digital subtraction angiography, most diagnostic radiographic procedures used 
biplane imaging with the occasional use of single plane imaging during the evaluation of 
aneurysms of neck origin. In therapeutic embolization procedures, on the other hand, 
single plane radiography may provide enough information to evaluate the progress of the 
embolization during the intermediate stages of vessel occlusion. Thus, embolization 
procedures made extensive use of single plane radiography. Biplane radiography was 
still required to make definitive evaluation of the embolization result at the more critical 
stages of the procedure. 



Imaging Techniques 

In fluoroscopy, the automatic brightness control (ABC) adjusts the x-ray tube 
voltage to yield the appropriate amount of light at the output of the image intensifier. 
Figure 3-1 shows the relative frequency at which different tube voltages were used during 
fluoroscopy of a typical interventional neuroradiologic procedure. Relative frequencies 
were computed by determining the fraction of surface dose delivered to the patient at 
each kVp interval. In the frontal plane, the tube voltages mostly used during fluoroscopic 
imaging were distributed between 66 kVp and 95 kVp, most frequently in the 81-85 kVp 
range. In the lateral plane, tube voltages between 61 kVp and 85 kVp were equally used 
during fluoroscopy with a more frequent use of the 71-75 kVp range. In general, the tube 
voltages used in the frontal plane were shifted about 10 kVp higher to those of the lateral 












37 

plane. The difference in physical thickness of the head region between frontal and lateral 
views explains such differences. 



30% 




56-60 



71-75 86-90 101-105 

X-Ray Tube Voltage (kVp) 



Figure 3-1 : Histogram of surface dose contribution at different x-ray tube voltages from 
frontal plane (black bars) and lateral plane (gray bars) fluoroscopy for an 
average interventional neuroradiologic procedure. 



In digital radiography, the tube voltage is determined from the associated 
fluoroscopic techniques. Figure 3-2 shows the relative frequency at which different tube 
voltages were used during radiography of a typical interventional neuroradiologic 
procedure. Similarly to fluoroscopy, the distribution of radiographic tube voltages in the 
frontal plane was shifted about 10 kVp higher to that of the lateral plane. The most 
frequently used voltages in the frontal plane were located at the 76-80 kVp range. Tube 
voltages at the 81-95 kVp range were also used extensively during radiography in the 



38 

frontal plane. In the lateral plane, voltages between 61-75 kVp and 86-90 kVp were most 
frequently used. The range of 66-70 kVp signifies the radiographic tube voltages 
primarily used in the lateral plane. 



30% 



(A 

ID 

Q. 
> 

* 20% 
o 

c 
o 

3 

o 



> 

1 



10% 



0% 



56-60 



■ Frontal Radiography 
□ Lateral Radiography 




71-75 86-90 101-105 

X-Ray Tube Voltage (kVp) 



Figure 3-2: Histogram of surface dose contribution at different x-ray tube voltages from 
frontal plane (black bars) and lateral plane (gray bars) radiography for an 
average interventional neuroradiologic procedure. 



The Patient Dosimetry System 
System Description 

A patient dosimetry system (PEMNET*) was installed in April 1995 on each of 
the two x-ray imaging planes of the Toshiba neurobiplane KXO-80C/D unit. The 



1 PEMNET: Patient Exposure Monitoring Network. Clinical Microsystems Inc., Arlington, VA. 



39 

PEMNET unit is a microprocessor-based system running its own on-board software. 
Eight units can be networked to a single PC server via RS-121 interfaces, through which 
they transfer patient dosimetric data to the PC server for storage and analysis, or receive 
calibration information from the PC. The PEMNET system does not measure surface 
doses directly, as may be the case of dose area product meters (Shrimpton and Wall, 
1982). Instead, the system is passively hardwired to the x-ray generator to acquire the 
input signals listed in Table 3-1. These input signals permit the computation of surface 
doses that patients would receive, if it were assumed that the same skin area is continually 
exposed to the x-ray beam. 

The PEMNET dosimetry system computed patient surface doses by using the x- 
ray tube radiation output at the selected technique factors (kVp and mA) together with 
information about the patient location relative to the x-ray tube and measured exposure 
times. The patient location was determined from the height of the x-ray table relative to 
the x-ray tube or by using an ultrasonic sensor at orientations where the position of the 
table was not relevant, as in lateral views. When the x-ray table intercepted the x-ray 
beam, x-ray attenuation by the table was taken into account. The surface dose was 
computed in digital radiography, whereas the surface dose rate was determined in 
fluoroscopy. In both digital radiography and fluoroscopy, the patient dosimetry system 
calculated surface dose rates by sampling the radiation technique factors every 5 ms, and 
by computing an average exposure rate every 800 ms. The surface skin exposure rate and 
the cumulative surface skin exposure were displayed in real time for each imaging plane 






40 

on two panel displays adjacent to the image display monitors and were readily visible by 
the neuroradiologic staff. 



Table 3-1: List of the Input Signals Interfaced to the PEMNET Dosimetry System from 

the Toshiba Neurobiplane Imaging Unit 

Signal Description Comments 



Tube Potential (kV) Radiographic or fluoroscopic 

Tube Current (mA) Radiographic or fluoroscopic 

Pulsed Fluoroscopy Current (mA) 20 mA or 50 mA 

Table Height (cm) Relative to the floor plane 

C-Arm Height (cm) Relative to the floor plane 

C-Arm Angulation (°) RAO and CAU rotations 

Ultrasonic Distance Measurement (cm) Active after a C-arm rotation of 1 5° 
NOTE : RAO ■ right anterior oblique; CAU = craniocaudal. 

Calibration 

The x-ray tube radiation output was dejgrmined from exposure measurements 
obtained using an MDH 1015C § radiation monitor with a 10x5-6 ionization chamber 
attached to the surface of an RSD RS-235" anthropomorphic head phantom as depicted 
in Figure 3-3. For exposure calibrations, the ionization chamber was located at the 
isocenter of each C-arm and in direct contact with the head phantom. For frontal (PA) 
exposures, the ionization chamber was located at the occipital area of the 



s Radcal Corporation, Monrovia, CA 

" Radiology Support Devices Inc; Long Beach, CA 



41 

anthropomorphic phantom, while for lateral exposures the chamber was located next to 
the temporal bone of the phantom. All measurements of entrance skin exposures 
included the contribution of backscatter radiation. The entrance skin exposure was 
converted to the surface dose using the expression 



87.7 






muscle 



D = — x ^ \" xjf mGy (3.1) 



10 



f/f, 
, P 



where D is dose to muscle in mGy for monoenergetic photons, and X is the exposure in 

roentgens; {Mjp) muscle is the mass energy absorption coefficient of muscle, and (juJpXir is 

the mass energy absorption coefficient of air. The ratio of mass energy absorption 

coefficients of muscle to air does not change significantly with energy (about 4% 

between 30 keV and 100 keV x rays) and can be taken to be equal to 1.06 for 

polyenergetic, diagnostic x-ray spectra (Johns and Cunningham, 1983; Jones and Wall 

1985; Wall et al., 1988). The dose D to muscle from polyenergetic x-ray spectra then is 

given by 

87 7 
Z> = ^-xl.06xA' = 9.30xZ mGy (3.2) 

10 

where both the dose in muscle and the exposure in air include contribution from 

backscatter radiation. 

X-ray generator signals fed to the patient dosimetry system were calibrated to read 

the correct technique factors, source-to-patient distance and tube orientation. The 

ultrasonic sensors attached to the side of each x-ray tube collimator were calibrated to 

measure the x-ray source-to-patient surface distance directly. Measured surface doses 

were entered into the patient dosimetry system and transferred to the PC server along 



42 

with kVp, mA, and exposure time information. A calibration program on the PC server 
generated corresponding surface dose curves (third and fourth degree polynomials) as a 
function of the applied kVp and mAs at different modes of operation (i.e., radiographic or 
fluoroscopic) and transferred the curve coefficients back to the system's microprocessor. 
Two calibrations were performed separately for each imaging plane, with and without the 
presence of the x-ray table, in order to derive the table attenuation coefficients. 





FIGURE 3-3: Calibration setup of the frontal plane (left) and lateral plane (right) using an 
RSD RS-235 anthropomorphic head phantom. 



Evaluation 

The accuracy of the patient dosimetry system was evaluated in all fluoroscopic 
and radiographic modes of operation. In fluoroscopy, the Toshiba neurobiplane unit may 
be operated either in continuous or pulsed (15 or 30 frames/sec) mode. In the 
radiographic mode, the unit may be operated either in cut film (CF) mode or in digital 
subtraction angiography (DSA) mode. Each acquisition mode was investigated using the 
geometry of a typical patient setup. The ionization chamber was attached to the 
anthropomorphic phantom as shown in Figure 3-3. For the frontal plane system 



43 

evaluation, table attenuation and positioning were taken into consideration. There was no 
table attenuation during testing of the lateral plane. The source-to-patient distance was 
measured directly by the ultrasonic sensor in the lateral plane. 

Measured surface doses were compared to the corresponding values computed by 
the patient dosimetry system for the experimental arrangements listed in Table 3-2. Table 
3-3 shows ratios of the measured, X M , to calculated, X c , surface doses obtained with the 
patient dosimetry system. Average XJX C ratios (± one standard deviation) over clinical 
kV and mAs ranges are given in Table 3-3, together with the total number of individual 
data points recorded. Changing the source to patient distance or the electronic 
magnification during continuous fluoroscopy resulted in an average XJX C ratio of 
1.04±0.03. Simulation of a non-standard examination performed in the frontal plane, 
with "maximized" changes made to all possible imaging parameters, resulted in a surface 
dose computed by the patient exposure system of 0.98 Gy whereas the measured value 
was 0.93 Gy (5% difference). 

In general, these results demonstrated that the patient dosimetry system would 
normally generate surface doses, which are within 5% of the true surface dose. The 
uncertainties of threshold radiation doses for the induction of deterministic effects such as 
skin erythema or epilation are considerably larger than 5% (Wagner et al., 1994; Rubin 
and Casarett, 1968; UNSCEAR, 1988) due to factors such as the anatomical location and 
size of the irradiated region, tissue vascularity and oxygenation, as well as the patient age, 
genetic background and hormonal status. Thus the accuracy of the patient dosimetry 
system is adequate for measuring surface doses to patients undergoing interventional 
neuroradiologic procedures. 



44 



Table 3-2: Experimental Arrangements for Evaluating the Patient Dosimetry System 



Arrangement 



Variables 



Purpose 



Imaging techniques 



Tube voltage 

Tube current 

Exposure time 

Frame rate 

Electronic magnification 

Geometric magnification 



Image Acquisition Modes 



Continuous fluoroscopy 
Pulsed fluoroscopy 

Cut film 

Digital subtraction 
angiography 



Complete 

neurointerventional 

procedure 

(-1.0 Gy) 



Image acquisition modes 
Imaging techniques 

Table height and location 

Source-to-image receptor 
distance 

C-arm rotation 



Evaluate the system 
response to technique 
changes 



Evaluate the system 
response in different image 
acquisition modes 



Simulate a complete patient 
examination maximizing 
changes which are 
technically possible to 
determine an upper limit of 
the accuracy of the system 



Collimation 



45 



Table 3-3: Summary of the Ratios of the Measured to Calculated Surface Doses, XJX C , 
Obtained During Testing of the Accuracy of the Patient Exposure System 



Operating 


Number 


Frontal 


Lateral 




Mode 


of Tests 


Plane 


Plane 




Continuous 
Fluoroscopy 


28 


0.99 ± 0.03 


1.04 ±0.02 




Pulsed 
Fluoroscopy 


13 


0.96 ± 0.02 


0.96 ± 0.02 


kV and mAs techniques 
were varied 


Radiography 
(CF and DSA) 


10 


0.94 ± 0.05 


1.01 ±0.01 





Continuous _ . , _ . , _, 

„ 5 mm 1.03 1.01 

r luoroscopy 



Pulsed 
Fluoroscopy 



10 min 



Radiography 70 

(CF and DSA) frames 



0.93 



0.93 



1.03 



1.00 



Automatic brightness 
control (ABC) was used 



NOTE : CF ■ cut film acquisition; DSA ■ digital subtraction acquisition 
SOURCE : Gkanatsios et ah, 1997. 



Data Acquisition 

Following the introduction of the patient dosimetry system into clinical practice, 
dosimetry data were obtained for 175 patients undergoing interventional neuroradiologic 
examinations. At the end of each patient examination, the recorded surface dose data 
were automatically uploaded to the PC server for subsequent analysis. A database with 
information shown in Figure 3-4 was built. Dosimetry data were analyzed to provide 
cumulative doses for each imaging mode on both imaging planes for the complete patient 
neurointerventional procedure. In addition, dosimetry data were also obtained for 
discrete kV intervals, as well as for discrete dose rate intervals. 



46 

Additional information made available by the patient dosimetry system included 
the total fluoroscopic time, the number of times fluoroscopy was engaged, and the total 
number of radiographic (cut film and DSA) images acquired. 



B PEMNET Log 



sE 



■ 



g a TO T Si li l! 



VTFB 



ROOM 12: Toshiba Biplan 



Wikst, hkmcJMoi 



Patient ID: 

Type of Exam: 

Date of Exam: 

Age: 



CeiebsalArteiioqram 



Biplane: ^ 
Malel^ FemaleT 



Attending Radiologist: 

Fellow: 

Resident: 



FRONTAL PLANE 



CONTINUOUS 
FLUOROSCOPY 



Time: 

Exposure: 
Engage: 

R<10: 
R<20: 
R>20: 



PULSED 
FLUOROSCOPY 



Time: 

Exposure: 

Engage: 

R<10: 
R<20: 
R>20: 



DIGITAL 
SUBTRACTION 



Exposure: 
Flames: 

R<10: 
R<20: 
R>20: 



<B0 kV 

60 kV 

65 kV 

70 kV 

75 kV 

80 kV 

85 kV 

90 kV: 

95 kV 

lOOkV 

105 kV 

110 kV 




<60 kV 

60 kV 

65 kV 

70 kV 

75 kV 

80 kV 

85 kV 

90 kV: 

95 kV 

100 kV: 

105 kV: 

110 kV 




<60 kV 

60 kV 

65 kV 

70 kV 

75 kV 

80 kV 

85 kV 

90 kV 

95 kV: 

100 kV 

105kV 

110 kV 



Record: i< ( < | |~ 




34 ► I >i !►«! of 114 



3 



3 



~3 



#4 ^ 



TOTALS 



Exposure: 



R<10: 
R<20: 
R>20: ) 



d 



FIGURE 3-4: Sample page from the PEMNET database showing all recorded information 
for the frontal imaging plane. 



47 

Fluoroscopy 
Dosimetric data including the surface dose received by the patient from use of 
fluoroscopy, the total time of fluoroscopy, and the rate at which dose was delivered to the 
patient were recorded by the patient dosimetry system. Additional recorded information 
included the number of times fluoroscopy was engaged and the x-ray tube voltages used 
in fluoroscopy during the course of a neuroradiologic procedure (seen in Figure 3-4). 



X-Rav Beam Localization 

During neurointerventional procedures, fluoroscopy was used to position the 
catheter next to the vessel anomaly in the brain in order to inject contrast and 
subsequently image the anomaly. Since the transfemoral artery technique was used to 
guide the catheter to the vertebral or carotid arteries, some fluoroscopy was performed 
over the truncal and thoracic regions of a patient. After studying the use of fluoroscopy 
for ten patients, it was determined that, on average, about thirty seconds (34 ± 1 1 sec) of 
frontal plane fluoroscopy were spent on the truncal region and an additional two minutes 
(133 ± 77 sec) at the upper thoracic, lower neck region. 

The amount of fluoroscopy performed over the truncal and thoracic regions was 
relatively independent of the patient and the type of neurointerventional procedure. 
Therefore, the surface dose corresponding to 2.5 minutes of fluoroscopy was subtracted 
from the dose contributed by use of frontal plane fluoroscopy. The remaining dose was 
considered to be absorbed in the head region of the patient. To subtract this fraction from 
the surface dose to the head, the average dose rate was computed for each patient, 



48 

multiplied by 2.5 minutes and subtracted from the total surface dose corresponding to 
frontal plane fluoroscopy. 

During interventional neuroradiologic procedures, a 20°-30° rotation of the x-ray 
source in the sagittal plane of the patient may be used when acquiring radiographic 
images. Although the central axis of the x-ray beam changes position on the surface of 
the head with rotation of the x-ray source, there are parts of the x-ray beam, which 
overlap before and after rotation. Such overlaps indicate that there are areas that will 
always be exposed to radiation regardless of the applied x-ray source rotation. Thus, any 
rotation of the x-ray source could be ignored when computing surface doses from 
radiographic exposures, since the maximum surface dose to any given area of the head is 
of interest. 



Surface Doses 

Figure 3-5 shows the histogram distribution of the patient surface doses received 
from fluoroscopy alone. The median values of the fluoroscopic surface doses were 0.32 
Gy and 0.11 Gy for the frontal and lateral imaging planes, respectively. Maximum 
surface doses were computed at 2.4 Gy for the frontal plane and 2.7 Gy for lateral plane. 
The data shown in Figure 3-5 do not differentiate between diagnostic and therapeutic 
procedures. 

The distribution of surface dose in frontal plane fluoroscopy was mainly spread 
over the range of 0.0-0.8 Gy. In the lateral plane, the majority of patients (70%) received 
less than 0.2 Gy with some patients (17%) receiving between 0.2-0.4 Gy. The lateral 
plane was mainly used for catheter position verification and less for catheter 



49 

manipulation, which kept the surface doses in the lateral plane low in comparison to the 
frontal plane. Surface doses at the tail of the dose distribution for each plane (above 0.6- 
0.8 Gy) corresponded to embolization neuroradiologic procedures. Such procedures 
require use of additional fluoroscopy for catheter positioning and verification at the site 
of occlusion. Twenty-seven (15%) out of 175 patients recorded underwent cerebral 
embolization. 



150 




FRONTAL 

Median = 0.32 Gy 
(Maximum) = 2.4 Gy 

LATERAL 

Median = 0.11 Gy 
(Maximum) = 2.7 Gy 



0.00-0.20 0.81-1.00 1.61-1.80 2.41-2.60 

Surface Absorbed Dose (Gy) 



FIGURE 3-5: Histogram distribution of surface doses for 175 patients from frontal plane 
(black bars) and lateral plane (gray bars) fluoroscopy. 



Surface Dose Rates 

Figure 3-6 shows the histogram distribution of the rate at which surface doses 
were delivered to the patient during fluoroscopy. The median values of the fluoroscopic 



50 

surface dose rates were 37 mGy/min for the frontal plane and 43 mGy/min for the lateral 
plane. The maximum skin dose rates recorded by the patient dosimetry system were 
approximately 100 mGy/min for both planes. Since patient thickness is smaller in the 
lateral dimension of the head, the automatic brightness control selects a lower tube 
voltage (also seen in Figure 3-1), which increases the surface dose rate. 

The histogram distribution of the surface dose rate in the frontal imaging plane 
presents a normal distribution shape, but is widely spread over the range of mGy/min to 
55 mGy/min. The dose rate distribution in the lateral plane is more concentrated at the 
16-30 mGy/min and 46-65 mGy/min. In general, fluoroscopic imaging may vary 
significantly from patient to patient due to variations in source-to-surface distance and the 
selection of imaging techniques (i.e., kVp/mA). 



30 



1 20 

'■5 

£ 



a> 
1 10 




FRONTAL 

Median = 37 mGy/min 

(Maximum) = 99 

mGy/min 

LATERAL 

Median = 43 mGy/min 

(Maximum) =101 

mGy/min 



0.0-5.0 30.1-35.0 60.1-65.0 90.1-95.0 

Surface Absorbed Dose Rate (mGy/min) 

Figure 3-6: Histogram distribution of surface dose rates for 175 patients from frontal 
plane (black bars) and lateral plane (gray bars) fluoroscopy. 



51 

Fluoroscopic Times and Intervals 

Other useful information recorded by the patient dosimetry system included the 
total time of fluoroscopy and the number of times fluoroscopy was engaged during an 
interventional neuroradiologic procedure. Figure 3-7 shows the histogram distribution of 
total fluoroscopic times in each imaging plane. The frontal plane was most frequently 
used with a median value of 12 minutes per patient examination compared to a median 
value of 3.0 minutes in the lateral plane. The fluoroscopic times in the frontal plane were 
more spread over the range of 5-20 minutes in comparison to the lateral plane, which 
used less than 5 min of fluoroscopy for the majority (73%) of the patients. The data 
shown in Figure 3-7 include both diagnostic and embolization neuroradiologic 
procedures with the higher values corresponding to the latter. For embolization 
procedures, the duration of fluoroscopy may be extended well beyond the median values 
to times as high as 70 minutes and 41 minutes for the frontal and lateral planes, 
respectively. 

Figure 3-8 shows a histogram distribution of the number of times fluoroscopy was 
engaged on each imaging plane. The median number of times that the operator initiated 
fluoroscopy was 62 in the frontal plane and 26 in the lateral plane. This difference clearly 
indicates the extensive use of fluoroscopy in the frontal imaging plane during 
interventional neuroradiologic procedures. Increased number of fluoroscopic instances 
also indicate higher surface doses (seen in Figure 3-5) and longer fluoroscopic times 
(Figure 3-7) between the frontal and lateral imaging planes. 



52 



150 



120 

B 

c 

0) 

S 90 

Q. 



5 60 

E 

3 



FRONTAL 

Median = 12 min 
(Maximum) = 70 min 

LATERAL 

Median = 3.0 min 
(Maximum) = 41 min 




r— r r ~ ' i — r r ~ t i i i 1 r - — i 1 r 

0.0-5.0 20.1-25.0 40.1-45.0 60.1-65.0 

Fluoroscopic Time (min) 



Figure 3-7: Histogram distribution of fluoroscopic times to 175 patients from frontal 
plane (black bars) and later plane (gray bars) fluoroscopy. 



Catheter positioning primarily done using the frontal imaging plane varies widely 
from patient to patient. This difference between frontal and lateral imaging planes 
introduced a wider spread to the histogram distribution of fluoroscopic intervals 
corresponding to the frontal plane, as shown in Figure 3-8. Both distributions, however, 
show long tails with a maximum of 226 fluoroscopic intervals in the frontal plane and 
170 intervals in the lateral plane. Such tails on the distribution may account for the need 
to use additional fluoroscopy during embolization procedures. 



53 



60 



W 

*-• 

C 

© 
3 

ra 
Q. 



45 



30 



a> 

E 

3 



15 

























FRONTAL 

Median = 62 
(Maximum) = 226 

LATERAL 

Median = 26 
(Maximum) ■ 1 70 



0.0-15.0 



90.1-105.0 180.1-195.0 

Fluoroscopic Intervals 



FIGURE 3-8: Histogram distribution of fluoroscopic intervals for 175 patients from 
frontal plane (black bars) and lateral plane (gray bars) fluoroscopy. 



Radiography 
Dosimetric data on the surface dose received by the patient from use of 
radiographic imaging (i.e., cut film and DSA images), the number of radiographic frames, 
and the dose per frame were recorded by the patient dosimetry system. Additional 
recorded information included the radiographic tube voltages used during the course of a 
neuroradiologic procedure (seen in Figure 3-4). 



X-Ray Beam Localization 

Radiographic image acquisitions are performed almost exclusively on the head 
region during interventional neuroradiologic procedures. The majority of them is 



54 

performed employing the digital subtraction angiography (DSA) technique. As discussed 
under x-ray beam localization for fluoroscopy, the main goal is to compute the maximum 
doses delivered to any surface of a patient's head. Similarly to fluoroscopy, radiography 
may also be considered unaffected by the small degree of x-ray source angulation usually 
applied to the frontal imaging plane in the sagittal plane of the patient. 



Surface Doses 

Figure 3-9 shows the histogram distribution of the patient surface doses received 
from radiographic acquisitions. The median values of the radiographic surface doses 
were 0.80 Gy and 0.50 Gy for the frontal and lateral planes, respectively. The maximum 
radiographic dose recorded in the frontal plane was 4.8 Gy, twice the maximum dose 
recorded in fluoroscopy for the same plane. In the lateral plane the maximum 
radiographic surface dose was 3.8 Gy, about 30% higher than the maximum dose from 
fluoroscopy in the same plane. 

Although the histogram distribution of the radiographic surface doses in the 
frontal plane has a longer tail and a higher median value than the dose distribution in the 
lateral plane, both distributions are very similar. This supports the fact that frontal and 
lateral plane radiography are equally utilized during any type of interventional 
neuroradiologic procedure. 



Surface Dose Rates 

Figure 3-10 shows the histogram distribution of the surface dose per frame in 
radiographic imaging with median values of 2.5 mGy/frame and 1.8 mGy/frame for the 



55 

frontal and lateral planes, respectively. Maximum doses of 5.6 mGy/frame were recorded 
in the frontal plane and 4.9 mGy/frame in the lateral plane. In general, the size and 
densities of the head region do not vary significantly among patients. Similar tube 
voltages would be used for all radiographic imaging acquisitions. Therefore, the 
distribution of doses per frame depends mostly on changes to the source-to-surface 
distance (i.e., use of different degrees of magnification among patients). In the frontal 
plane where geometric magnification is more frequently used, the dose per frame 
distribution approaches that of a wide normal shaped distribution. In the lateral plane 
where magnification is not uses as often, the distribution is steeper (less variability). 



75 



I 50 

(0 

£ 










-■ 


—1 












FRONTAL 

Median = 0.80 Gy 
(Maximum) = 4.8 Gy 

LATERAL 

Median = 0.50 Gy 
(Maximum) = 3.8 Gy 



0.00-0.30 1.21-1.50 2.41-2.70 3.61-3.90 
Surface Absorbed Dose (Gy) 

Figure 3-9: Histogram distribution of surface doses for 175 patients from frontal plane 
(black bars) and lateral plane (gray bars) radiography. 



56 



Radiographic Frames 






Figure 3-11 shows the histogram distribution of the number of radiographic 
(DSA) frames acquired during diagnostic and therapeutic neuroradiologic procedures. 
Median values of 353 frames and 316 frames were recorded for the frontal and lateral 
plane, respectively. Due to the complexity of some embolization procedures, however, 
the number of frames acquired to evaluate the progress of an occlusion may run as high 
as 1400 in the frontal plane and 1000 in the lateral plane. 




FRONTAL 

Median = 2.5 mGy/frame 
(Maximum) = 5.6 mGy/frame 

LATERAL 

Median =1.8 mGy/frame 
Maximum) = 4.9 mGy/frame 



0.00-0.30 1.51-1.80 3.01-3.30 4.51-4.80 

Surface Absorbed Dose per Frame (mGy/frame) 

FIGURE 3-10: Histogram distribution of surface dose rates for 175 patients from frontal 
plane (black bars) and lateral plane (gray bars) radiography. 



As shown by the distribution of radiographic frames in both planes, the number of 
imaging frames required in each procedure is variable. Depending on the degree of 
difficulty of extracting diagnostic information from the acquired images, as well as the 



57 

type of anomaly to be diagnosed, the number of radiographic images acquired is normally 
between 100 and 500 frames. Embolization procedures, on the other hand, may require 
two to four times the number of radiographic images to complete the associated tasks. 



45 



FRONTAL 

Median = 353 frame 
(Maximum) = 1388 frame 

LATERAL 

Median = 316 frame 
(Maximum) = 999 frame 




0-100 401-500 801-900 1201-1300 

Number of Radiographic Frames 

FIGURE 3-11: Histogram distribution of the number of radiographic frames for 175 
patients from frontal plane (black bars) and lateral plane (gray bars) 
radiography. 



Conclusions 
Figure 3-12 shows the histogram distributions of the total surface dose to the 
patient from the use of fluoroscopy and radiography during an interventional 
neuroradiologic procedure. The medians of the total surface dose were 1 .2 Gy and 0.64 
Gy for the frontal and lateral plane, respectively. The maximum surface dose received by 
a patient was of the order of 5 Gy for both imaging planes. The majority of the doses 



58 

were concentrated between 0.2 Gy and 1 .2 Gy for both imaging planes. A significant 
number of higher doses, however, was indicated by the tails of the two histogram 
distributions. Most of the dose was contributed by radiography, which accounts for the 
67% of the total surface dose in the frontal plane and 78% of the total dose in the lateral 
plane. Fluoroscopy only accounted for the 33% and 22% of the total surface dose in the 
frontal and lateral plane, respectively. 

Although 28% of the patients in this study may have exceeded the nominal 
threshold absorbed dose to the skin for the induction of deterministic effects (2 Gy), there 
were no cases of epilation or skin erythema observed in this series of 175 patients. This 
is not surprising since any erythema would be fleeting and faint. Epilation would only be 
identified by a slightly different amount of hair loss, as perceived when combing one's 
hair, and would not require total loss of hair. For acute radiation exposures, observable 
effects such as total epilation are more likely to occur at doses in excess of 6 Gy (Huda 
and Peters, 1994). 

Several factors need to be considered in predicting the likelihood of deterministic 
effects to patients undergoing neuroradiologic examinations. One factor is the fact that 
radiation doses are delivered over an extended time period, which may be as long as 
several hours. Of great importance is also the fact that the radiation field is varied over 
the patient. For individuals with the highest radiation exposures, the neuroradiologist 
generally makes a concerted effort to either move the relative orientation of the x-ray 
beam or to utilize the orthogonal imaging plane in so far as these options do not adversely 
impact the required imaging information. Many of the neuroradiologic imaging 
procedures also make use of wedge shaped transparent filters which serve to reduce the 



59 

radiation doses at the periphery of the x-ray field of view whilst maintaining image 
quality within the central region. All these factors reduce the likelihood of deterministic 
injuries to patients and should be practiced during extended neuroradiologic procedures. 



60 



J 40 

re 
£ 



o 

Si 

I 20 







; i—i 






FRONTAL 

Median = 1.2 Gy 
(Maximum) = 5.0 Gy 

LATERAL 

Median = 0.64 Gy 
(Maximum) = 5.2 Gy 

*W~~W P i ""IF P P* i ri 



0.00-0.30 1.51-1.80 3.01-3.30 4.51-4.80 

Surface Absorbed Dose (Gy) 

Figure 3-12: Histogram distribution of the total surface doses to 175 patients from 
frontal plane (black bars) and later plane (gray bars) fluoroscopy and 
radiography combined. 



CHAPTER 4 
ENERGY IMPARTED AND EFFECTIVE DOSE IN NEURORADIOLOGY 



Introduction 

The effective dose, E, is a dosimetric parameter which takes into account the doses 
received by all irradiated radiosensitive organs and may be taken to be measures of the 
stochastic risk (ICRP, 1977, 1991). Although the effective dose is an occupational dose 
quantity based on an age profile for radiation workers, this dose descriptor is being 
increasingly used to quantify the amount of radiation received by patients undergoing 
diagnostic examinations which use ionizing radiation (ICRP, 1987; NCRP, 1989; 
UNSCEAR, 1993). Notwithstanding the fact that there are problems associated with 
converting effective doses to a corresponding detriment (Huda and Bews, 1990), there are 
important benefits to be gained by using effective dose to quantify patient doses in 
diagnostic radiology. One advantage is that the effective dose attempts to measure the 
stochastic risk to the patient, which is the motivation for all patient dosimetry studies in 
diagnostic radiology. In addition, the effective dose to a patient undergoing any 
examination may be compared to that of any other radiologic procedure as well as natural 
background exposure and regulatory dose limits (ICRP, 1991; NRC, 1995a, 1995b). 

Measurements or computations of effective doses from x-ray examinations are 

difficult and time consuming. An additional problem is that most measurements or 

calculations make use of a standard phantom based on the reference man as defined by 

the International Commission on Radiological Protection (ICRP, 1975). Although the 

60 



61 

importance of patient size for medical radiation dosimetry has been recognized 
(Lindskoug, 1992; Chappel et al., 1995), it is not obvious how to scale the effective dose 
computed for standard man to different size patients, such as pediatric patients, who 
undergo similar examinations. These limitations impede the wider use of effective dose 
in radiology. Huda and Gkanatsios (1997) developed a more practical approach to 
compute effective doses from energy imparted for a variety of radiologic examinations 
and different size patients including infants and children. This method was used in this 
chapter to compute effective doses from computed values of energy imparted to patients 
undergoing interventional neuroradiologic procedures. 

Exposure-area products to different regions of the body and at different tube 
voltages were used to compute values of energy imparted, e, from interventional 
neuroradiologic procedures (Gkanatsios, 1995; Gkanatsios and Huda, 1997). Values of 
energy imparted were converted to patient effective dose, E, using Ele conversion factor 
corresponding to the projections and body regions irradiated during interventional 
neuroradiologic procedures. Values of Ele for the posterio-anterior (PA) projections of 
the abdomen, chest and cervical spine and for the PA and lateral (LAT) views of the head 
were obtained from radiation dosimetry data computed using Monte Carlo calculations on 
an adult anthropomorphic phantom (Hart et al, 1994a). This method was extended to 
determine effective doses to pediatric patients who differ in mass from the adult sized 
phantom. 



62 

Method 



Energy Imparted 

The energy imparted, s, to a patient undergoing any radiologic x-ray examination 
can be estimated by modeling the phantom as a slab of water with thickness z using the 
expression 

s=coxESExA J (4.1) 

where co is the energy imparted per entrance exposure-area product, ESE is the exposure 
measured free-in-air at the beam entrance plane of the phantom, and A is the area of 
exposure also measured at the entrance plane (Gkanatsios, 1995; Gkanatsios and Huda, 
1997). 

The parameter co depends on the water phantom thickness z, the x-ray tube 
voltage and x-ray beam half-value layer (Gkanatsios, 1995; Gkanatsios and Huda, 1997) 
Values of co can be computed from 

co=ax HVL + J3 J R/' cm 2 (4.2) 

where a and /? are coefficients that depend on tube voltage and phantom thickness, and 
HVL is the half-value layer of the x-ray beam at a given tube voltage in mm of 
aluminum. Figure 4-1 shows the behavior of co as a function of water phantom thickness 
for x-ray tube voltages of 60 kVp, 80 kVp and 100 kVp as apply to the x-ray tube and 
table filtration of the neuro-biplane Toshiba imaging system. Examples of a and fi 
coefficients and half-value layers of the x-ray beams at different tube voltages are given 
in Table 4-1. 



63 



200 



150 - 



• 100kVp(HVL = 5.3mmAI) 
x 80 kVp (HVL = 4.3 mm Al) 
A60kVp(HVL = 3.2mmAI) 




1 < ' ' 



' ' 



10 20 30 

Water Phantom Thickness (cm) 



40 



FIGURE 4-1 : Values of a; as a function of water phantom thickness for tube voltages of 
60 kVp, 80 kVp and 100 kVp. 

NOTE : The values of co were computed for constant voltage 
waveforms, an x-ray tube anode angle of 11° and 6.0 mm Al 
filtration (x-ray tube filtration plus table filtration of the Toshiba 
frontal imaging plane). 



The free-in-air entrance exposures to the patient, ESE, were obtained from the 
patient exposure data recorded by the frontal and lateral patient exposure meters at 5 kVp 
intervals (Figure 3-4). The recorded exposures included the contribution of backscatter 
radiation from an RSD RS-235 n anthropomorphic head phantom. Therefore, backscatter 
fractions measured using the same phantom were subtracted from the recorded exposures. 
Table 4-2 lists measured backscatter fractions for the RSD RS-235 anthropomorphic head 
phantom as a function of tube voltage for the frontal and lateral imaging planes. 



n Radiology Support Devices Inc; Long Beach, CA 



2.64 


1.85 


3.23 


2.23 


3.77 


2.59 


4.32 


2.96 


4.85 


3.35 


5.34 


3.74 


5.80 


4.12 


6.23 


4.51 



64 



Table 4-1 : Computed a and /? Coefficients and Half- Value Layers for X-Ray Beams as a 

Function of Tube Voltage 

Tube Voltage _ _ . _ _, _ . HVL— Frontal HVL— Lateral 

,. XT : or Coefficient B Coefficient , ... . ... 

(kVp) 2 (mm Al) (mm Al) 

50 2.275E-05 1.300E-05 

60 2.229E-05 1.895E-05 

70 2.147E-05 2.521E-05 

80 2.031E-05 3.215E-05 

90 1.899E-05 3.910E-05 

100 1.771E-05 4.557E-05 

110 1.654E-05 5.145E-05 

120 1.549E-05 5.673E-05 

NOTE : a and /? coefficients were computed for a water phantom 
thickness of 20 cm. The half- value layers were determined for constant 
voltage waveforms, an x-ray tube anode angle of 11°, 6.0 mm Al 
filtration for the frontal imaging system (x-ray tube filtration plus table 
filtration) and 3.0 mm Al filtration for the lateral imaging system (x-ray 
tube filtration). 



Energy imparted values were computed separately for fluoroscopy and 
radiography. In frontal plane fluoroscopy, the abdominal, upper thoracic, lower neck and 
head regions were irradiated. As determined in Chapter 3, about thirty seconds of frontal 
plane fluoroscopy were spent on average on the abdominal region and an additional two 
minutes at the upper thoracic, lower neck region. The exposures corresponding to these 
fluoroscopic times were used to compute the energy imparted to the abdomen and upper 
chest, lower neck body regions. The remaining fluoroscopic exposure was focused over 
the head region and was used to compute the energy imparted to the head. In lateral 



65 

fluoroscopy, frontal radiography and lateral radiography all exposure was taken to be 
incident on the head region. 

Table 4-2: Backscatter Fractions of Radiation Exposure at Different Tube Voltages 



Tube Voltage 


Backscatter 


Backscatter 


(kVp) 


Factor (Frontal) 


Factor (Lateral) 


50 


0.056 


0.119 


60 


0.073 


0.130 


70 


0.084 


0.135 


80 


0.096 


0.140 


90 


0.107 


0.143 


100 


0.110 


0.147 


110 


0.116 


0.154 


120 


0.121 


0.157 


NOTE: Backscatter fractions were determined u; 


;ing the RSD RS-235 



anthropomorphic head phantom. 



For the purpose of computing energy imparted, the water equivalent thickness of 
the irradiated region as well as the area of exposure at the x-ray beam entrance surface 
were required. Table 4-3 lists the water equivalent thickness and exposure area of the 
head regions corresponding to different age groups used to compute energy imparted to 
the head. Table 4-4 lists the water equivalent thickness and exposure area for different 
age groups used to compute energy imparted to the abdominal and upper thoracic, lower 
neck regions. 



66 



Table 4-3: Patient Thickness and Area of Exposure Corresponding to the Head Region 

of Different Age Groups 



Patient 

Age 


Head 
Density 

(g/cm 3 ) 


PA 

Thickness 

(cm) 


LAT 

Thickness 

(cm) 


PA Area of 

Exposure 

(cm 2 ) 


Newborn 


1.057 


12.3 


9.51 


86.7 


1-yr-old 


1.071 


16.7 


13.1 


160 


5-yr-old 


1.090 


19.8 


15.5 


221 


10-yr-old 


1.095 


20.6 


16.2 


239 


1 5-yr-old 


1.104 


21.6 


17.2 


265 


Adult 


1.112 


22.2 


17.8 


279 




NOTE: PA thickness 


and LAT thickness r 


epresent the equiv 


alent 



thickness of a water phantom computed from the physical dimensions 
and density of the head. The area of exposure for each patient group in 
the PA projection was computed using the physical dimensions of the 
head. The area of exposure in the LAT projection was estimated to be 
equivalent to 1 .2 of the corresponding areas in the PA projection. 
SOURCE : Densities and physical diameters of the head region at 
different age groups were taken from Huda et al., 1997. 



Adult Effective Doses 

The National Radiological Protection Board (NRPB) have performed a 
comprehensive series of Monte Carlo dose calculations for the most common x-ray 
projections {Hart et al, 1994a). The Monte Carlo runs made use of a hermaphrodite 
anthropomorphic phantom with a mass of 70.9 kg and a height of 174 cm, which included 
the female breasts, ovaries, uterus and testes. Each Monte Carlo run tracked the pattern 
of energy deposition in the anthropomorphic phantom from primary and scattered 
photons for total 4,000,000 photons used with each x-ray projection. A total of 68 
separate views were obtained using x-ray spectra generated between 50 kVp and 120 kVp 



67 

with added filtration ranging from 2 mm Al to 5 mm Al. X-ray spectral data were 
obtained using an updated version of a computer program published by lies (1987). 



Table 4-4: Patient Thickness and Area of Exposure Corresponding to the Trunk Region 

of Different Age Groups 



Patient 
Age 



Trunk 
Density 

(g/cm 3 ) 



Abdomen 


Chest/Neck 


Thickness 


Thickness 


—PA— 


— PA— 


(cm) 


(cm) 


9.75 


9.00 


13.0 


11.2 


15.0 


13.2 


16.9 


13.8 


20.2 


14.6 


20.4 


15.0 



PA Area of 

Exposure 

(cm 2 ) 



Newborn 

1-yr-old 

5-yr-old 

10-yr-old 

1 5-yr-old 

Adult 



0.995 
1.002 
1.000 
1.005 
1.030 
1.018 



175 
175 
175 
175 
175 
175 



NOTE : The PA thickness represents the equivalent thickness of a water 

phantom computed from the physical diameters and density of the 

trunk. 

SOURCE : Densities and physical diameters of the trunk region at 

different age groups were taken from Huda et al., 1997. 



For each x-ray examination, the Monte Carlo dosimetry data generated by the 
NRPB permitted the computation of the effective dose, E, as defined by the International 
Commission on Radiological Protection (ICRP, 1977, 1991). The phantom breast dose 
and the mean of the testes and ovary doses were used to determine the contributions to 
the effective dose from the breast and gonads, respectively. The Monte Carlo dosimetry 
data also provided the mean doses to three body regions consisting of the head, D h , trunk, 



68 

D„ and legs, D,. Mean doses to these three body regions were used to compute the mean 
energy imparted to the patient, e, using the equation 

e = D h x 5.8 + D, x 43.0 + D, x 22.1 J (4.3) 

where the mass** of the head is 5.8 kg, the mass of the trunk, including the arms, is 43.0 
kg and the mass of the legs is 22.1 kg. 

The complete dosimetry results of these Monte Carlo simulations have been made 
available in a software format (Hart et al, 1994b) and were used to obtain the values of 
effective dose and energy imparted for specific projections as applied to radiation 
exposures during interventional neuroradiologic procedures. These projections were the 
posterio-anterior projections of the abdomen, chest, cervical spine, and head regions, as 
well as the right lateral projection of the head region. For each x-ray projection, values of 
Ele were computed at eight tube voltages ranging between 50 kV and 120 kV and 
generated at 10 kV intervals with a beam filtration equivalent to 3.0 mm aluminum 
(lateral plane) and 6.0 mm aluminum (frontal plane). The effective dose per unit energy 
imparted, Ele (mSv J" 1 ), for the projections of the trunk and head regions are given in 
Table 4-5. The average Ele ratios of the chest and C-spine projections at each kVp were 
used to determine effective doses from irradiation of the upper thoracic, lower neck 
region. 



n Wall BF. Private communication (1996). 



69 

Pediatric Effective Dose 

By definition, 1 Gy of uniform whole body irradiation to x-rays results in an 
effective dose of 1 Sv and is independent of the mass of the exposed individual. For a 
70.9 kg anthropomorphic adult subject to uniform whole body irradiation, energy 
imparted can be directly converted into effective dose with one joule corresponding to an 
effective dose of 14.1 mSv. For uniform whole body irradiation, the effective dose E(M) 
to an individual with a mass M (Table 4-6) who absorbs a total of s J is given by 

E(M) = £xl4.\x— mSv (4.4) 

M 

Figure 4-2 shows how the effective dose varies with the patient mass for uniform whole 
body irradiation with a total of one joule imparted to the individual. 

For nonuniform exposures normally encountered in diagnostic radiology, the 
relative radiosensitivity of the irradiated region needs to be taken into account when 
obtaining the effective dose. The relative radiosensitivity of any body region remains 
approximately constant with age (ICRP, 1991; Almen and Mattsson, 1996). For instance, 
if the head accounts for x% of the total stochastic risk in adults uniformly exposed to x- 
rays, this body region will also account for approximately x% of the total stochastic risk 
for any other age group. As a result, the effective dose to a patient of mass M kg for a 
given x-ray projection i who absorbs s joules of energy is obtained using 

E(M) = eJ^\ xl^- mSv (4.5) 



70 

where (E/s) i is the ratio of effective dose to energy imparted (mSv J" 1 ) obtained for the 
same projection i in the adult anthropomorphic phantom with a mass of 70.9 kg. 
Standard masses of patients of different ages are given in Table 4-6. 



Table 4-5: Values of Effective Dose per Unit Energy Imparted, Ele in mJ/Sv, for 
Different Body Projections as a Function of Tube Voltage 



kVp 



Abdomen 
(PA) 



Chest 
(PA) 



C-Spine 
(PA) 



Head 
(PA) 



Head 
(LAT) 



50 


10.7 


60 


12.0 


70 


13.2 


80 


13.9 


90 


14.5 


100 


14.9 


110 


15.4 


120 


15.8 




NOTE: Th< 



12.8 


4.19 


13.6 


4.67 


14.1 


5.04 


14.6 


5.40 


15.0 


5.66 


15.4 


5.92 


15.6 


6.11 


15.8 


6.27 



4.06 
4.62 
5.00 
5.40 
5.68 
5.90 
6.17 
6.32 



4.08 
4.56 
4.94 
5.29 
5.61 
5.87 
6.06 
6.24 



NOTE : The values of Ele corresponding to PA views were computed 
for 6.0 mm Al filtration (frontal imaging plane). The values of Els 
corresponding to the Head LAT view were computed for 3.0 mm Al 
filtration (lateral imaging plane). 



Age 
Group 



Patient 
Mass (kg) 



Table 4-6: Standard Patient Mass for Different Age Groups 



SOURCE : Huda et al, 1997. 



Newborn 1-yr-old 5-yr-old 10-yr-old 15-yr-old Adult 
3.4 9.8 19 32 55 70.9 



71 



1000 



> 
to 

E, 

in 

of 

8 100 

Q 

i 

LU 



10 



*< 



3.0 kg for a new born 



70.9 kg reference man- 



1 ' i i i i i_ 



10 
Patient Mass (kg) 



100 



Figure 4-2: Effective dose as a function of patient mass for one joule of uniform whole 
body irradiation. 



Adult Patient Doses 
The following sections summarize the energy imparted and effective doses to adult 
patients from interventional neuroradiologic procedures. One hundred and forty nine 
adult patients, 132 of them underwent diagnostic angiographic and seventeen underwent 
therapeutic embolization procedures, were studied. Fluoroscopy and radiographic 
acquisitions were reviewed separately. 



Energy Imparted 

Figure 4-3 shows the histogram distribution of the energy imparted to patients 
from use of fluoroscopy during interventional neuroradiologic procedures. The median 



72 

value of energy imparted was 1 .77 J with energy imparted in the frontal plane being the 
major component of fluoroscopy. The distribution of energy imparted in fluoroscopy was 
mainly spread over the range of 0-5 J. Fourteen (9%) of 149 adult patients received more 
than 5 J with three (2%) patients receiving more than 10 J of energy imparted from 
fluoroscopy with a maximum value of 12.6 J. Although there was no separation done in 
the distribution between diagnostic angiographic and therapeutic embolization 
procedures, the median value of energy imparted to patients who underwent 
embolizations was 3.48 J. Six of the seventeen embolization patient exceeded the value 
of 5 J. 



60 



S 40 




0-1 



BIPLANE FLUOROSOPY 

Median =1.77 J 
Maximum = 12.6 J 

FRONTAL FLUOROSOPY 

Median =1.33 J 
Maximum = 9.67 J 



LATERAL FLUOROSOPY 

Median = 0.31 J 
Maximum = 8.98 J 



4-5 8-9 

Energy Imparted (J) 



12-13 



Figure 4-3: Histogram distribution of energy imparted to patients from use of 
fluoroscopy during interventional neuroradiologic procedures. 



73 

Figure 4-4 shows the histogram distribution of the energy imparted to patients 
from radiographic acquisitions during interventional neuroradiologic procedures. The 
median value of energy imparted was 4.30 J. The distribution of energy imparted from 
radiographic acquisitions was mainly spread over the range of 0-10 J. Sixteen (1 1%) of 
149 adult patients received between 10 J and 15 J. A maximum value of 21.2 J was 
recorded. The median value of energy imparted to patients who underwent therapeutic 
embolization procedures was 8.40 J. Seven of the seventeen embolization patients 
received energy imparted from radiographic acquisitions greater than 10 J. 



30 



w 

| 20 

« 
O- 



0) 

E 

z 



10 







- 


















] 
I 


BIPL 
FRO! 
.ATE 


A> 

Ma 

VT 

IV 
Ma 

y 

Ma 


IE RADIOGRAPHY 

ledian = 4.30 J 
Lximum = 21.2 J 

AL RADIOGAPHY 

Median = 2.71 J 
iximum= 16.6 J 

LL RADIOGRAPHY 

[edian= 1.26 J 
ximum= 10.9 J 



0-1 



5-6 10-11 15-16 

Energy Imparted (J) 



20-21 



FIGURE 4-4: Histogram distribution of energy imparted to patients from radiographic 
acquisitions during interventional neuroradiologic procedures. 



Figure 4-5 shows the histogram distributions of the total energy imparted to adult 
patient from diagnostic angiographic and therapeutic embolization neuroradiologic 



74 

procedures. The median value of the total energy imparted was 6.69 J. The maximum 
energy imparted received by a patient was 26.9 J. The majority of the patients who 
underwent interventional neuroradiologic procedures received up to 14 J of energy 
imparted. Fifteen (10%) of the adult patients shown by the tail of the distribution in 
Figure 4-5 received energy imparted between 14 J and 27 J. The median value of the 
total energy imparted to patients who underwent therapeutic embolization procedures was 
13.3 J. Eight of the seventeen embolization patients exceeded the 14 J value of total 
energy imparted. The largest fraction of energy imparted was produced by radiographic 
acquisitions. The average fraction of energy imparted from radiographic acquisitions was 
about 66% of the total energy imparted. Only one third of the total energy imparted was 
accounted for use of fluoroscopy. 



Effective Doses 

Figure 4-6 shows the histogram distributions of the total effective dose to adult 
patient from diagnostic angiographic and therapeutic embolization neuroradiologic 
procedures. The median value of the total effective doses was 36 mSv. The majority of 
the patients who underwent interventional neuroradiologic procedures received between 
10 mSv and 70 mSv of effective dose. The tail of the histogram distribution shown in 
Figure 4-6 corresponds to nineteen (13%) patients who received effective doses greater 
than 70 mSv. The median value of the total effective dose to patients who underwent 
therapeutic embolization procedures was 74 mSv. Ten of the seventeen embolization 
patients exceeded the 70 mSv value of total effective dose. As in surface doses and 
energy imparted, radiographic acquisitions accounted for the largest fraction of the 



75 

effective dose to adult patients. On average, about 64% of the effective dose 
corresponded to radiographic acquisitions. Use of fluoroscopy accounted for only one 
third of the total effective dose received by patients during interventional neuroradiologic 
procedures. 




0-2 



8-10 16-18 

Energy Imparted (J) 



24-26 



FIGURE 4-5: Histogram distribution of the total energy imparted to patients undergoing 
diagnostic angiographic and therapeutic embolization neuroradiologic 
procedures. 



Pediatric Patient Doses 
The following sections summarize the energy imparted and effective doses to 
pediatric patients from interventional neuroradiologic procedures. Twenty-six pediatric 
patients (younger than 20 years of age), sixteen of them underwent diagnostic 



76 

angiographic and ten underwent therapeutic embolization procedures, were studied. 
Fluoroscopy and radiographic acquisitions were reviewed separately. 




0-10 



BIPLANE 
FLUOROSCOPY & RADIOGRAPHY 

Median = 36 mSv 
Maximum = 156 mSv 

FRONTAL PLANE 

Median = 24 mSv 
Maximum ■ 1 02 mSv 



LATERAL PLANE 

Median = 8.3 
Maximum = 94 mSv 



60-70 120-130 

Effective Dose (mSv) 



Figure 4-6: Histogram distribution of the total effective dose to patients from biplane 
neuroradiologic examinations. 



Energy Imparted 

Figure 4-7 plots the energy imparted to pediatric patients from fluoroscopy as a 
function of patient mass. The mass of each patient was interpolated from Table 4-6 
according to the age of the patient. As Figure 4-7 shows, there is no significant 
correlation of energy imparted to patient mass. The median value of energy imparted 
from fluoroscopy to all pediatric interventional neuroradiologic procedures was 1.04 J. 
Pediatric patients who underwent therapeutic embolization procedures had a median of 



77 

1 .62 J. The median value of energy imparted from fluoroscopy to adult patients who 
underwent interventional neuroradiologic procedures was 1.77 J. 



I 



O) 

LU 







- 






BIPLANE FLUOROSOPY 


- 


■ 




■ Therapeutic 


.: 




X Diagnostic 


- 




■ 


X 

■ 


X 


■ 


t 2 = 0.02 


■ X 

x x 


x- — x- — 




■ 


— 1 1 1 1 1 1 1 u. 


X 

■ 1 1 1 1 1 1 


X 

1 ' ' ' i ' 


x * x 

— 1 — 1 1 1 1 1 1 1 1—1 L_J l_l 1 1 1 1 







20 40 60 

Patient Mass (kg) 



80 



Figure 4-7: Energy imparted as a function of patient mass from fluoroscopy during 
interventional neuroradiologic procedures on pediatric patients. Line 
shows the linear fit between energy imparted and patient mass. 



Figure 4-8 plots the energy imparted to pediatric patients from radiographic 
acquisitions as a function of patient mass. As was the case for fluoroscopy, there was no 
significant correlation between the energy imparted from radiographic acquisitions and 
patient mass. The median value of energy imparted from radiographic acquisitions to all 
pediatric interventional neuroradiologic procedures was 2.01 J. Pediatric patients who 
underwent therapeutic embolization procedures had a median value of energy imparted of 



78 

2.61 J. The median value of energy imparted from radiographic acquisitions to adult 
patients who underwent interventional neuroradiologic procedures was 4.30 J. 



15 



12 



•a 

® ~ 

r 9 

re 

a. 

E 

en 6 

c 
• 

c 







BIPLANE RADIOGRAPHY 

■ Therapeutic 
X Diagnostic 



x 



x 



^ = 0.17 




20 40 60 

Patient Weight (kg) 



80 



Figure 4-8: Energy imparted as a function of patient mass from radiographic 
acquisitions during interventional neuroradiologic procedures on 
pediatric patients. Line shows the linear fit between energy imparted and 
patient mass. 



Figure 4-9 plots the total energy imparted to pediatric patients during 
interventional neuroradiologic procedures as a function of patient mass. As Figure 4-9 
shows, there was no significant correlation between total energy imparted from 
interventional neuroradiologic procedures and patient mass. The median value of total 
energy imparted to all pediatric interventional neuroradiologic procedures was 3.45 J. 
Pediatric patients who underwent therapeutic embolization procedures had a median 
value of energy imparted of 4.09 J. The median value of energy imparted from 



79 

radiographic acquisitions to adult patients who underwent interventional neuroradiologic 
procedures was 6.69 J. 



£U 










FLUOROSCOPY 












& 












RADIOGRAPHY 


cr- 15 - 












3, 










■ Therapeutic 


0) 

r. 

m 






■ 




x Diagnostic 


o- .^ 










■ 


E 10 - 












>» 






1 




X 


TO 










X 


5 






■ 






e 


- 






r 2 = 


0.13 . 


m 5 










- — ■ X 




X 


X 


■ 


X 


■ I 


- 


> 

_l l_J I 


J I I 


1 

1 1 1_1 1 1 1 1 1—1 1 | 1 L 


X 


X 

I 1 ! | 1 1 1 1 1 1 1 I 1 







20 40 60 

Patient Mass (kg) 



80 



FIGURE 4-9: Energy imparted as a function of patient mass from interventional 
neuroradiologic procedures on pediatric patients. Line shows the linear 
fit between energy imparted and patient mass. 



Effective Doses 

Figure 4-10 plots the total effective dose to pediatric patients from interventional 
neuroradiologic procedures as a function of patient mass. Although the pediatric data of 
the total effective doses are widely scattered (r 2 = 0.3), a linear correlation between 
effective dose and patient mass is evident. The median value of total effective dose to all 
pediatric interventional neuroradiologic procedures was 44 mSv and was higher 
compared to the median value of 36 mSv effective dose to adult patients. Pediatric 



80 

patients who underwent therapeutic embolization procedures had a median value of 66 
mSv effective dose. 



zlUU 












FLUOROSCOPY 














& 


_ 200 - 


X 










RADIOGRAPHY 


S" 














CO 

E 












■ Therapeutic 
















co 150 - 












X Diagnostic 


W 
O 






■ 








Q 














■ 


- 


X 










£ 100 - 


: 




■ 


■ 






o 


■ 










Adult median 


0) 














St 












■ \ 


LU 

50 


! i 

i 




x~~~~^^ 


-^w = 


= 0.30 


*\s 










1 


X 


"^""""""""^Srr 





_j i i 


J I 1- 


— 1 l—l 1 i 1 1 


—1 1— I 1 


X 

J 1 1 1 


x 1 

—j — i — i i i i . i i i_ 



20 40 60 

Patient Mass (kg) 



80 



Figure 4-10: Effective dose as a function of patient mass from interventional 
neuroradiologic procedures on pediatric patients. Line shows the linear 
fit between effective dose and patient mass. 



Discussion 
Major errors in determining energy imparted to patients result when estimating the 
equivalent water phantom thickness, z, and due to the implicit differences between a 
(finite) heterogeneous patient and a semi-infinite homogeneous water phantom. Figure 
4-1 shows the energy imparted per unit exposure-area product, a>, as a function of 
phantom thickness for a range of x-ray tube voltages. The largest increase of co with 
phantom thickness is expected at small thicknesses, given that the mean free path of 



81 

monoenergetic photons in water ranges from 4.4 cm at 50 keV to 6.6 cm at 140 keV. 
Once the phantom thickness reaches about three or four mean free paths, most of the x- 
ray photons will have been absorbed and any further increase of the phantom thickness 
will have little affect on co. 

Figure 4-1 shows that at 80 kV, the thickness of the water phantom used to simulate 
a patient for the purposes of estimating energy imparted generally will not be a critical 
parameter for applications with phantom thicknesses greater than 15 cm. Since the water 
equivalent size of an adult head is between 1 8 cm (lateral view) and 22 cm (frontal view), 
small deviations from the average sizes given in Table 4-3 will have a minimal effect on 
the computation of energy imparted. A difference between a 20 cm and 22 cm phantom 
thickness at 80 kVp is of the order of 2%. Even for pediatric patients where the size of 
the head is smaller (13 cm to 17 cm for 1-yr-olds), a 2 cm error in estimating the water 
equivalent thickness will result in a maximum error of about 5% when computing energy 
imparted. 

Minor errors in computing energy imparted to patients arise from the use of 
diverging x-ray beams in clinical applications and the presence of nonuniformities in x- 
ray beam intensity due to the heel effect. The former is likely to be of negligible 
importance whereas the latter could easily be accounted for by experimentally obtaining 
an average entrance skin exposure over the beam area. Measuring the exposure at the 
centerline of the x-ray beam is also a good approximation of the average exposure over 
the entire field. Another error in determining energy imparted from irradiation to the 
head region results from occasional use of wedge shaped transparent filters which serve 
to reduce the radiation doses at the periphery of the x-ray field of view whilst maintaining 



82 

image quality within the central region. Such filters are used during imaging of the 
frontal view of the head and can attenuate the entrance exposure by about 50% at 80 kVp. 
As these filters cover an area between 10% and 20%, an overestimate of the energy 
imparted from frontal imaging plane fluoroscopy of the order of 5% to 10% can occur. 

Use of Equation (4.4) permits the determination of the approximate values of 
effective doses to pediatric patients who undergo radiologic examinations. The NRPB 
has recently published dosimetric data on pediatric patients ranging from the newborn to 
15 year olds (Hart et al, 1996). Figure 4-11 shows a comparison between the Els values 
obtained using Equation (4.4) (continuous line) with the NRPB data (solid circles), which 
were determined by performing Monte Carlo calculations in a range of anthropomorphic 
phantoms of different age. Differences between these two data sets, when averaged over 
the five ages investigated, were 17% with the largest differences shown for the 1-yr-old 
(31%) and 5-yr-old (36%) phantoms. Such differences may be due to pediatric heads 
accounting for a markedly larger fraction of the total body masses in these ages compared 
to adults. It is of interest to note, however, that use of different types of anthropomorphic 
phantoms to determine pediatric effective doses in planar radiography can result in 
differences in effective dose of the order of 30% (Hart et al, 1996b). 

In general, the effective doses computed in this work compare three to six times 
higher to values published by others for similar interventional neuroradiologic procedures 
(Feygelman et al, 1992; Bergeron et al, 1994; McParland, 1998; Berthelson and 
Cederblad, 1991). However, all other reported values refer to limited number of 8-28 
procedures, and none of them made use of means of recording radiation exposures in real 
time. Different imaging equipment, setup and imaging procedures among institutions 



83 

play a major role to how different effective doses may be among institutions. The fact 
that Shands hospital at the University of Florida is an academic institution that trains new 
neurointerventional radiologists may also account for some of the differences between the 
recorded effective doses in this work and others. 



Conclusions 
Values of energy imparted from interventional neuroradiologic procedures were 
high due to the demands and complexity of these procedures. The median value of the 
total energy imparted to adult patients who underwent interventional neuroradiologic 
procedures was 6.69 J. Pediatric patients received a median value of energy imparted of 
3.45 J from interventional neuroradiologic procedures. In the case of therapeutic 
embolization procedures, additional use of fluoroscopy is required for catheter 
manipulation and positioning at the site of occlusion, as well as extensive radiographic 
acquisitions to evaluate the progress of the occlusion. Such demands increased the 
median values of energy imparted to adults undergoing therapeutic embolization 
procedures to 13.3 J. Pediatric embolizations received a median value of 4.09 J. Overall, 
radiographic acquisitions accounted for two thirds of the total energy imparted, with 
fluoroscopy contributing only one third. There was no significant correlation between 
energy imparted from interventional neuroradiologic procedures and patient mass. 






84 



1000 



^ 100 
> 

V) 

E 






10 -: 



Equation (4.4) 



• Newborn 



1-yr 






_j i i . . 



5-yr 10 . yr 

15-yr 



H "- 

10 
Patient Mass (kg) 



_l ' ' ' l_l L 



100 



FIGURE 4-1 1 : Comparison of Els values vs. patient age as determined by Equation (4.4) 
and by using the dosimetry data from Hart et al. (1996a). 

NOTE : Values of Els were computed for the right lateral 
projection of the head. 



Effective doses computed for the 149 adult patients who underwent interventional 
neuroradiologic procedures had a median value of 36 mSv. Pediatric patients received a 
median effective dose of 44 mSv from interventional neuroradiologic procedures. The 
median effective dose to adults undergoing therapeutic embolization procedures was 74 
mSv. Pediatric embolizations received a median effective dose of 66 mSv. As was the 
case for energy imparted, radiographic acquisitions accounted for two thirds of the total 
effective dose, with fluoroscopy contributing only one third. Unlike energy imparted, 
effective doses showed a good linear correlation with patient mass. 



85 

The use of the effective dose permits an estimate of stochastic risk to be obtained 
by using current stochastic risk coefficients (ICRP, 1991; UNSCEAR, 1993; NAS, 1990). 
At the last attempt of the ICRP (1991) to estimate absolute stochastic risks from whole- 
body irradiation, a risk coefficient of 5xlO" 5 cancers per mSv of effective dose was 
derived. Using this risk coefficient, the median effective dose of 36 mSv to adult patients 
would result in one fatal cancer for every 555 interventional neuroradiologic procedures. 
An effective dose of 74 mSv to adults undergoing therapeutic embolization procedures 
would result in one fatal cancer for every 270 such procedures. The immediate, life 
saving benefits of interventional neuroradiologic procedures, however, far outweigh the 
risk of distant stochastic effects associated with these procedures. Also, such risk 
coefficients need to be treated with great caution given the current uncertainties 
associated with the extrapolation of radiation risks from high doses to those normally 
encountered in diagnostic radiology (Fry, 1996; Puskin and Nelson, 1996). 

In the case of pediatric patients undergoing interventional neuroradiologic 
procedures, it is important to note that any resultant stochastic detriment will depend on 
the age of the exposed individual. The stochastic radiation risks of carcinogenesis and 
genetic effects are generally greater for children than for adults to at least a factor of two 
(ICPR, 1991; NCRP, 1985). These factors would need to be taken into account when 
converting any pediatric effective doses into a value of risk or detriment. As a result, 
direct comparisons of pediatric doses with those of adults need to be treated with 
circumspection. 



CHAPTER 5 
IMAGE QUALITY 



Image Acquisition 
A phantom made of acrylic incorporating 1 .0 mm diameter cylindrical vessels 
filled with iodinated contrast was used to investigate signal detection during digital 
subtraction angiography (DSA). The detection of signal from iodinated vessels was 
evaluated by studying the threshold iodine contrast concentration detected in images 
acquired under specified parameters using digital image subtraction. 



Phantom Description 

Figure 5-1 illustrates the phantom used to simulate 1.0 mm diameter vessels for 
the purpose of evaluating image quality in neuroradiology. The phantom consists of 
stacked acrylic blocks with dimensions of 30 cm x 30 cm x 1 .3 cm. An insert holder 
made of acrylic with a thickness of 1 .3 cm is positioned at the center of the phantom to 
accommodate a vessel insert. A blank and a vessel insert measuring 30 cm x 9.0 cm x 
1 .3 cm were made out of acrylic. The blank insert was used to acquire mask images 
during digital subtraction angiography. The vessel insert had thirty cylindrical vessels 
1.0 mm in diameter and 35 mm in length drilled along its midplane at intervals of 8.0 mm 
apart. The total phantom thickness was 16.5 cm of acrylic, which was taken to be 



86 



87 

equivalent to about 20 cm of water taking the density of acrylic to be 1.19 g/cm 3 (Shleien, 
1992). 



Acrylic 
Holder 




Insert Slide 

Circle defines 
imaging area 



Vessel 
Insert 



Blank 
Insert 



Acrylic 
Blocks 







































1 




-1* 








1 
































FIGURE 5-1: Schematic diagram of the acrylic phantom with the vessel and blank inserts 
used to simulate small vessels for the purpose of evaluating image quality 
in neuroradiology. 



88 

The vessels on the acrylic insert were filled with iodinated contrast prepared from 
Ultravist §§ 300 iopromide solution diluted in heparin solution. The iodine concentrations 
in the contrast medium used to fill each vessel ranged from 50 mg/cc iodine in contrast 
solution to about 5.0 mg/cc as given in Table 5-1. The iodine concentration in each 
vessel was made to be 92% of the previous concentration. 



Acquisition of Digitally Subtracted Images 

The general experimental setup shown in Figure 5-2 used an x-ray source to 
image receptor distance (SID) of 105 cm (maximum SID) with the acrylic phantom 
positioned so that the geometric magnification of the vessel insert was 1.2. A 10x5-6 
ionization chamber of an MDH 101 5C*" exposure meter was attached to the beam 
entrance surface of the phantom to record entrance exposure. A 10x5-60 ftt ionization 
chamber of a second MDH 1015C exposure meter was attached to the surface of the 
image intensifier behind the grid to record the input exposure to the image receptor. Both 
ionization chambers were positioned carefully not to overlap with the vessels of the 
vessel insert as shown in Figure 5-3. 

The 23 cm diameter image intensifier mode was used for all image acquisitions. 
The generator was set to manual techniques allowing fine adjustments of the tube voltage 
(kVp), tube current (mA) and exposure time (ms). The optical gain was electronically 
adjusted by changing the diameter of the iris located between the image intensifier output 
phosphor and TV camera lens to produce a constant video level. All digital subtraction 



u Berlex Laboratories, Wayne, NJ 
'** Radcal Corporation, Monrovia, CA. 



89 

angiography acquisitions were performed using the 0.6 mm focal spot size. The window 
and level of the displayed images were adjusted to optimize signal detection during each 
DSA image acquisition, so that the resulting images would not be contrast limited. 

Table 5-1: Iodine Contrast Concentration in Each Vessel of the Vessel Insert 



i- 
<u 
A 

s 

s 
Z 

"3 

(A 
V) 

s 

> 


15 


15.6 


O 

o. 

■ 

r> 

o 
a 

■ 

M 

n 

o 
a 
n 
n 
a 
it- 
a 
■ 

o 

3 

3 

M 

o 
n 


a 


£ 

a 

m 
c 

3 


s 

c 
o 
U 

1 

c 

a 
o 

u 

<u 

i 
■3 

o 

M 


14.3 


16 


< 

H 
■ 

e 

3 

a- 
n 


14 


16.9 


13.2 


17 


13 


18.4 


12.1 


18 


12 


20.0 


11.1 


19 


11 


21.7 


10.3 


20 


10 


23.6 


9.43 


21 


9 


25.7 


8.68 


22 


8 


27.9 


7.99 


23 


7 


30.3 


7.35 


24 


6 


33.0 


6.76 


25 


5 


35.8 


6.22 


26 


4 


38.9 


5.72 


27 


3 


42.3 


5.26 


28 


2 


46.0 


4.84 


29 


1 


50.0 


4.45 


30 



NOTE : The iodine concentration in each vessel was made to be 92% of 
the previous concentration. 



tTt The 10x5-60 ionization chamber has a gain of xlO for better resolution. 



90 




FIGURE 5-2: Experimental setup for DSA acquisitions. 





Figure 5-3: Position of the two ionization chambers relative to the vessel insert (left). 
Subtracted image (right). 



91 

The tube voltage was adjusted on the generator and monitored using a Machlett 
Dynalyser III*** with a digital display. The input exposure to the image intensifier was set 
by adjusting the product of tube current and exposure time (mAs). The input exposure 
was verified by the ionization chamber attached to the surface of the image intensifier. 
The optical gain was controlled by electronic adjustments of the iris diameter to produce 
a constant video level during all DSA image acquisitions. 

The effects of tube voltage, image intensifier input exposure and geometric object 
magnification were studied by varying the appropriate parameters. The following setup 
was used for each selected parameter: 

Tube voltage (kVp) . Baseline techniques were set at a tube voltage of 70 kVp, 
where the input exposure was adjusted to 120 uR/frame and the electronic iris was 
adjusted to produce a mean pixel value between 2000-2100. Then the tube voltage was 
varied from 50 kVp to 100 kVp and the input exposure was adjusted appropriately to 
maintain an average pixel value between 2000 and 2100, as shown in Table 5-2. To 
repeat the experiment at a higher image intensifier input exposure, the baseline 
techniques were reset at a tube voltage of 70 kVp, where the input exposure was adjusted 
to 440 uR/frame and the electronic iris was adjusted to produce a mean pixel value 
between 2000 and 2100. Then the tube voltage was varied from 50 kVp to 100 kVp and 
the input exposure was adjusted to maintain the same average pixel value, as shown in 
Table 5-2. 



m Greenwich Instrument CO., Inc., Greenwich, CT. 



92 



Table 5-2: Imaging Techniques During Tube Voltage Experiments 



kVp 



120 uR/frame 



440 uR/frame 



mAs 



II Expos Video 
(uR/frame) Level 



mAs 



II Expos Video 
(uR/frame) Level 



50 


28.0 


204 


2075 


80.0 


56 


14.4 


170 


2018 


47.5 


60 


9.28 


152 


2104 


33.0 


66 


5.40 


130 


2028 


20.5 


70 


4.00 


119 


2068 


17.0 


80 


2.24 


115 


2054 


7.50 


90 


1.50 


116 


2117 


4.60 


100 


0.960 


113 


2079 


3.20 



709 
623 
564 
505 
444 
405 
388 
382 



1914 
2052 
2031 
2085 
2018 
2056 
2080 
2075 



Image intensifier input exposure . The input exposure to the image 
intensifier was varied from about 50 uR/frame to 1000 pR/frame. A tube voltage 
of 70 kVp was used during all DSA image acquisitions. The optical gain was 
adjusted accordingly to maintain an average pixel value between 2000 and 2100 
for all selected image intensifier input exposures, as such pixel values correspond 
to clinical practice. 

Geometric object magnification. The position of the phantom was varied 
to achieve a range of geometric object magnification from 1.15 §ss to 2.0. A tube 
voltage of 70 kVp was used during all DSA image acquisitions. The image 
intensifier input exposure was first set to 120 pR/frame and the experiment was 



93 

repeated at 440 pR/frame. The mAs was adjusted accordingly to produce a mean 
pixel value between 2000 and 2100 as shown in Table 5-3. 



Table 5-3: Imaging Techniques During Geometric Object Magnification 

Experiments 



Mag 



120 uR/frame 



440 pR/frame 



mAs 



II Exp 



Video 



(pR/frame) Level 



mAs 



II Exp 



Video 



(pR/frame) Level 



1.15 


4.40 


120 


2100 


16.3 


1.20 


4.60 


119 


2068 


17.0 


1.40 


4.80 


118 


2008 


17.8 


1.60 


5.00 


116 


2036 


18.3 


1.80 


5.40 


122 


2000 


18.8 


2.00 


5.80 


118 


2072 


19.3 



446 
444 
440 
435 
435 
434 



2100 
2018 
2000 
2010 
2000 
2000 



During all digital subtraction acquisitions, a mask of a single frame was acquired 
at 3 frames/sec acquisition rate using the blank phantom insert. The blank insert was then 
replaced by the vessel insert and twenty additional frames were acquired with no frame 
integration at the same acquisition rate of 3 frames/sec. The tenth frame of each DSA 
image acquisition was always used for image evaluation. The beam entrance exposures 
to the phantom and to the image intensifier were recorded during each DSA acquisition 
sequence. 



m Geometry limited. 



94 
Dosimetry and Image Quality 



Dosimetry 

The exposure to the phantom was measured at the x-ray beam entrance plane 
using a 10x5-6 ionization chamber of an MDH 1015C. The measured exposure included 
backscatter radiation coming from the acrylic phantom. The entrance exposure per frame 
was obtained from the integral exposure measured for each image acquisition sequence 
divided by the number of acquired frames. A conversion factor of 2.58X10" 4 C/kg (1R) 
corresponding to an absorbed dose of 9.3 mGy for muscle tissue was used to convert the 
beam entrance exposure to surface dose as was shown in Equation 3-2. 

The energy imparted was computed using Eqs. 4-1 and 4-2 for the applied tube 
voltage and corresponding half-value layer. The water equivalent phantom thickness 
needed to determine energy imparted was taken to be 20 cm for all energy imparted 
computations. The exposure was obtained from direct exposure measurements during 
each image acquisition sequence and included backscatter. The backscatter radiation 
fraction was measured for the corresponding phantom and applied tube voltages and was 
subtracted from the measured exposures to obtain the free-in-air exposure. The exposure 
area at the beam entrance plane of the phantom was computed from geometry assuming 
that the beam area at the image intensifier plane was a circle with a diameter of 23 cm. 
The diameter and shape of the x-ray beam area were verified using film. 



95 

Image Quality Evaluation 

For the purpose of evaluating signal detection of small iodinated vessels during 
digital subtraction angiography, the threshold iodine contrast concentration was 
determined from the subtracted images of the vessel insert. An independent observer was 
used for this task. The criteria used to evaluate signal detection by the observer are given 
in Table 5-4 where a scale of 1-5 was defined to characterize the visibility of a given 
iodine concentration. 

The observer was trained on the criteria of scoring each iodine concentration 
using ten subtracted images similar to those asked to evaluate. All the images were then 
presented in a random order to the trained observer. For each image, the observer 
identified the first vessel that was not visible and scored the consecutive vessels up to the 
first vessel that was perfectly visible. The lowest concentration that was assigned a score 
of three was taken to be the threshold iodine concentration of each image. 

Table 5-4: Score Describing the Visibility of Each Iodine Contrast Concentration 



Score Comment 



1 Iodine concentration was not resolved 

2 Iodine concentration was barely resolved 

3 Iodine concentration was resolved to a confidence level of 50% 

4 Iodine concentration was well resolved 

5 Iodine concentration was perfectly resolved 



96 

Precision of Measurements 

A set of technique parameters, which delivered an exposure of approximately 300 
pR/frame to the input phosphor and produced a video level of 2000-2100 at 70 kVp, was 
used to acquire five digital subtraction acquisition sequences. These sequences were 
acquired at different stages of the experiment. The five image acquisition sequences were 
used to determine the measurement precision of the computed dosimetric parameters, as 
well as the precision of the reader performance when evaluating the threshold contrast 
concentration of the subtracted images. 

The exposure to the phantom was measured at the x-ray beam entrance plane and 
the surface doses and energy imparted were computed for each acquisition. The average 
entrance exposure measured was 97.6 ± 0.7 mR/frame. The average video level was 
2046 ± 9. An average surface dose of 0.908 ± 0.006 mGy/frame and an average energy 
imparted of 2.39 ± 0.02 mJ/frame were computed from the five image acquisition 
sequences. 

The subtracted images of the five acquisition sequences used to determine the 
precision were presented to the observer in a random order mixed with other similar 
subtracted images. The standard deviation of the five readings was taken to be the 
precision of the threshold contrast concentration and was computed to be ±12% of the 
threshold contrast concentration (8.0 ± 0.96 mg/cc). 



97 
Results 



Tube Voltage 

Table 5-5 (119 p.R/frame at 70 kVp) and Table 5-6 (444 p.R/frame at 70 kVp) 
summarize the results of surface dose, energy imparted and threshold iodine contrast 
concentration for varying tube voltage under constant video level during digital 
subtraction angiography. 

Figure 5-4 shows the surface dose and energy imparted as a function of tube 
voltage. To both low and high input exposures, the surface dose decreased by about 80% 
and the energy imparted dropped by almost 70% as the tube voltage increased from 50 
kVp to 100 kVp. The largest decrease in doses occurred between 50 kVp and 66 kVp. 

Figure 5-5 shows the threshold iodine concentration as a function of tube voltage. 
The threshold concentration increased faster with tube voltage at 120 p.R/frame compared 
to the 440 p.R/frame. At 120p.R/frame, the threshold iodine concentration was 
proportional to the kVp 201 , and increased about 100% when the tube voltage increased 
from 50 kVp to 70 kVp and another 60% from 80 kVp to 100 kVp. At 440 uR/frame, the 
threshold concentration became less sensitive to tube voltage changes and was 
proportional to kVp 156 . At this input exposure, changes to the tube voltage from 50 kVp 
to 70 kVp increased threshold iodine concentration by about 70%. Changes to the tube 
voltage from 80 kVp to 100 kVp increased threshold contrast by another 40%. 



98 



Table 5-5: Tube Voltage Dependency at 120 (iR/frame 



II Input Video Level f Energy 

kVp Exposure (mean pixel , _, .. . Imparted 

, nie v , x (mGy/frame) , .,. . 

(|aR/frame) value) J ' (mJ/frame) 



Threshold 

Iodine 

Concentration 

( m S /cc ) 



50 
56 
60 
66 
70 
80 
90 
100 



204 
170 
152 
130 
119 
115 
116 
113 



2075 


1.22 


2018 


0.811 


2104 


0.615 


2028 


0.439 


2068 


0.377 


2054 


0.279 


2117 


0.236 


2079 


0.196 



1.55 
1.17 
0.950 
0.739 
0.657 
0.543 
0.496 
0.438 



8.0 
11 
13 
12 
18 
28 
30 
30 



Table 5-6: Tube Voltage Dependency at 440 p.R/frame 



II Input Video Level Energy 

kVp Exposure (mean pixel , _ M . Imparted 

, nle x . ~Z (mGy/frame) , _% . 

(uR/frame) value) J (mJ/frame) 



Threshold 

Iodine 

Concentration 

( m s /cc ) 



50 
56 
60 
66 
70 
80 
90 
100 



709 
623 
564 
505 
444 
405 
388 
382 



1914 


3.94 


2052 


2.95 


2031 


2.36 


2085 


1.79 


2018 


1.34 


2056 


0.955 


2080 


0.653 


2075 


0.656 



5.03 
4.26 
3.65 
3.02 
2.33 
1.86 
1.37 
1.47 



6.2 
6.8 
8.7 
8.8 
10 
12 
17 
17 



99 



10.0 



0) 

E 

t 

CD 

E 

Q) 1.0 

i 
o 

Q 
e 

i 

CO 



0.1 



- 


120 iiR/frame 


A 


- 


•'"--A. 

\ "A-... 


-- 


v *"* - 

^y. a 


"*< A 


•^ 


- 


a Energy Imparted 


^^\« 


• Surface Dose 

1 | , ; ^_^_ 


j i ^—^ i i 



10.0 



m 

3 
CD 

1 

3 

■o 
fi 

- o i 

1.0 ® 

Q. 



| 

3 

CD 



0.1 



40 



60 



80 



100 



120 



Tube Voltage (kVp) 



10.0 



o 

E 
2 

o 

I 

•" 1.0 

(A 
O 

Q 



8 

i 

3 
(O 



0.1 



a Energy Imparted 
• Surface Dose 



440 iiR/frame 




_i ' ' 



4- 



J I L. 



-I I l_ 



10.0 



m 

3 
(D 

(2 
3 

fi> 

1.0 g 



3 

<D 



0.1 



40 



60 80 100 

Tube Voltage (kVp) 



120 



Figure 5-4: Surface dose and energy imparted as a function of tube voltage. 



100 



50 
40 

30 



o 

E 

G 20 
_o 

">^ 
CO 



c 

0) 

u 

c 
o 
o 

a> 

_c 

'■£ 
o 



8 
7 
6 

5 




40 



50 60 70 80 90 100 

Tube Voltage (kVp) 



FIGURE 5-5: Threshold iodine concentration as a function of tube voltage. The circles 
correspond to the 120 uR/frame and have been fitted to kVp 201 . The 
squares correspond to the 440 uR/frame and have been fitted to kVp [S6 . 



Image Intensifier Input Exposure 

Table 5-7 summarizes the results of surface dose, energy imparted and threshold 
iodine contrast concentration for varying input exposure to the image intensifier under 
constant video level and tube voltage during digital subtraction angiography. 

Figure 5-6 shows the surface dose and energy imparted as a function of image 
intensifier input exposure at a tube voltage of 70 kVp. At a given tube voltage, both 
parameters are directly proportional to the input exposure. For the particular DSA 
phantom and geometry used, the surface dose per 100 p.R/frame input exposure to the 
image intensifier was about 0.310 mGy. The energy imparted for the same 100 p.R/frame 
input exposure was about 0.540 mJ. 



101 



6.0 



^4.0 

1 

a> 

i 

I 

8 2.0 

i 

w 



0.0 



a Energy Imparted 
• Surface Dose 




6.0 



250 500 750 1000 

Image Intensifier Input Exposure (ft R/frame) 



FIGURE 5-6: Surface dose and energy imparted as a function of image intensifier input 
exposure at 70 kVp. 



Figure 5-7 shows the threshold iodine concentration as a function of image 
intensifier input exposure. In general, as the input exposure increases, the threshold 
concentration decreases. The data below 250 u.R/frame were fitted separately and 
showed to be proportional to D'° 59 , where D is the input exposure. Such proportionality 
indicates a behavior very similar to the theoretical quantum limited curve plotted as 
D'° - 50 . The data above 250 u.R/frame showed proportionality to Z)" 027 . At input exposures 
below 250 u.R/frame, increasing input exposure by a factor of two improved threshold 
concentration by 34%. An improvement of only 17% could be achieved by doubling the 
input exposure after 300 pR/frame. 



102 



Table 5-7: Image Intensifier Input Exposure Dependency 



Input Expos 
(uK/framc 


ure 

') 


Mean Video 

Level 
(pixel value) 


Surface Dose 
(mGy/frame) 


Energy 

Imparted 

(mJ/frame) 


Threshold 

Iodine 

Concentration 

(mg/cc) 


51 




2016 


0.172 




0.299 


28 


72 




2060 


0.232 




0.404 


26 


92 




2092 


0.296 




0.516 


22 


119 




2068 


0.377 




0.657 


18 


164 




2032 


0.506 




0.882 


14 


196 




2040 


0.598 




1.04 


13 


262 




2036 


0.908 




1.38 


12 


303 




2052 


0.903 




1.57 


11 


343 




2012 


1.026 




1.79 


11 


444 




2018 


1.336 




2.33 


10 


568 




2084 


1.72 




2.99 


9.4 


764 




2076 


2.35 




4.09 


8.7 


990 




2080 


3.03 




5.28 


8.7 




NOTE: A constant tube 
intensifier input exposures 


voltage of 70 


kVp 


was used for all 


image 



Geometric Object Magnification 

Table 5-8 (119 (iR/frame) and Table 5-9 (444 fiR/frame) summarize the results of 
surface dose, energy imparted and threshold iodine contrast concentration for varying 
geometric magnification under constant video level and tube voltage during digital 
subtraction angiography. 



103 



100 



u 

| 

E 



■fc 10 
C 
• 
o 

§ 

o 

0) 

c 

'•B 
o 



1 



-0.59 




Theoretical quantum limited region 
(D -o. 50) 



D 



0.27 



J I I I I I I L. 



__^^^^^__^_^__ 



10 100 1000 

Image Intensifier Input Exposure ((aR/frame) 

FIGURE 5-7: Threshold iodine concentration as a function of image intensifier input 
exposure for a constant video level at 70 kVp. 

Figure 5-8 shows the surface dose and energy imparted as a function of geometric 
object magnification at 120 p.R/frame and 440 (aR/frame. Surface dose increased as the 
square of increase in geometric magnification due to its inverse proportionality to the 
square of the source-to-surface distance. On the other hand, energy imparted remained 
almost constant with geometric magnification. The 10% increase in energy imparted 
across the magnification range shown in Figure 5-8 was caused by the increased in 
imaging techniques to maintain a constant video level (Table 5-3). The increase in 
radiographic techniques was caused by the reduced scatter reaching the image intensifier 
by moving the phantom farther away from the image intensifier to increase 
magnification. 



104 



Table 5-8: Geometric Object Magnification Dependency at 120 p.R/frame 

Geom II Input Video Level _ _ _ Energy , .. 

~,. 17 / • , Surface Dose T * J , Iodine 

Object Exposure (mean pixel , __ .. . Imparted _ 

I** / n ,r x ix (mGy/frame) , ¥ ,, . Concentration 

Mag ((aR/frame) value) v J ' (mJ/frame) 



1.15 


120 


1.20 


119 


1.40 


118 


1.60 


116 


1.80 


122 


2.00 


118 



2100 
2068 
2008 
2036 
2000 
2072 



0.320 
0.364 
0.531 
0.745 
0.960 
1.36 



0.638 
0.656 
0.674 
0.688 
0.711 
0.757 



13 
18 
12 
10 
8.7 
10 



Table 5-9: Geometric Object Magnification Dependency at 440 u.R/frame 



Geom II Input 
Object Exposure 
Mag (^R/frame) 



1.15 


446 


1.20 


444 


1.40 


440 


1.60 


435 


1.80 


435 


2.00 


434 



Threshold 
Iodine 



Video Level _ „ _ Energy 

, . , Surface Dose T t*. 

(mean pixel , __ „, „ Imparted 

, "\ (mGy/frame) . 7L . Concentration 
value) J ' (mJ/frame) , , „ 
. (mg/cc) 



2100 
2018 
2000 
2010 
2000 
2000 



1.18 
1.35 
1.96 
2.71 
3.56 
4.65 



2.27 
2.35 
2.39 
2.42 
2.42 
2.46 



10 
10 
9.4 
8.7 
6.2 
6.2 



105 



1.6 



m 

E 12 
c 

>» 

o 

E 

<d 0.8 
■ 
o 
Q 

8 

#0.4 

3 
V) 



0.0 



5.0 



CD 


4,0 


£ 




2 




M~ 








>» 




o 


3.0 


E 




■*«-■* 




CD 




(0 




o 

Q 


2.0 


o 




o 




ra 




t 




3 


1.0 



0.0 



120 nR/frame 




a Energy Imparted 

• Surface Dose 

— i — ■ — ■ — i — | — i — i — i — i — i — | — i i i i i i i i ' ' ' 



1.6 



m 

3 

8 
1.2 | 

3 

■o 

8 

0.8 | 



0.4 1 

CD 



0.0 



1.3 1.6 1.9 2.2 

Geometric Object Magnification 



440 |aR/frame 




a Energy Imparted 

• Surface Dose 

-j — i — i — i — | — i — i i i i_ 



5.0 

m 

4.0 5 

3 
^3.0-g 

a. 

CD 

a 

r 2.0? 

c_ 

I 

1.0 o 



0.0 



1.3 1.6 1.9 2.2 

Geometric Object Magnification 



Figure 5-8: Surface dose and energy imparted as a function of image intensifier input 
exposure at 70 kVp. 



106 

Figure 5-9 shows the threshold iodine concentration as a function of geometric 
object magnification at 120 uR/frame and 440 p.R/frame. The difference in threshold 
concentration between the 120 uR/frame and 440 uR/frame input exposures was about 
30% across the magnification range shown in Figure 5-9. The relationship between 
magnification and threshold iodine contrast was approximately linear. An increase by a 
factor of two to geometric magnification decreased threshold concentration by half. 



24 



o 
.o 

|» 18 

e 
o 







120uR/frame 




440 uR/frame 



J I I I—I L. 



J 1 1 I I I- 



24 



- 18 



- 12 







1.0 1.2 1.4 1.6 1.8 2.0 

Geometric Object Magnification 



2.2 



FIGURE 5-9: Threshold iodine concentration as a function of geometric object 
magnification at 70 kVp. 



107 
Discussion 



Patient Surface Dose 

The effects of tube voltage, input exposure and geometric magnification on 
surface dose were demonstrated in Figure 5-4, Figure 5-6 and Figure 5-8, respectively. A 
comparison of the changes introduced to the surface dose by varying these parameters to 
achieve a given change in threshold iodine concentration is given in Table 5-10 and Table 
5-11. The changes in surface dose listed in Table 5-10 have been computed using 70 
kVp, 119 pR/frame and xl.2 magnification as the starting point. The changes in surface 
dose given in Table 5-11 have been computed using 70 kVp, 444 pR/frame and xl.2 
magnification as the starting point. 

Figure 5-10 plots the changes in surface dose introduced by the tube voltage, 
input exposure and geometric magnification versus the corresponding changes in 
threshold iodine concentration. In general, the smallest increase in surface dose for the 
same decrease in threshold concentration can be achieved by lowering the tube voltage. 
At 120 pR/frame of input exposure where quantum mottle is the primary source of noise 
(Figure 5-7), magnification becomes the most inefficient parameter to improve signal 
detection. For achieving the same improvement in threshold concentration, 
magnification increased surface dose twice as much compared to input exposure and 
three to four times compared to tube voltage. In the case of 440 pR/frame input 
exposure, however, where quantum mottle does not appear to be the primary source of 
noise, input exposure becomes the most inefficient parameter to improve signal detection. 



108 



Table 5-10: Comparison of the Effects of Tube Voltage, Input Exposure and Geometric 
Magnification on the Surface Dose for a Range of Changes in Threshold Iodine 

Concentration at 120 jaR/frame 



Change in 

Threshold 

Concentration 


Change in SD with 
kVp 


Change in SD with 
II Exposure 


Change in SD with 
Magnification 


0.0% 


0.0% 


0.0% 




0.0% 


{17.3 mg/cc) 
-10% 


{70kVp) 
14% 


{119 /jR/frame) 
19% 




{xl.2) 
49% 


{15.6 mg/cc) 
-20% 


{67kVp) 
32% 


{142 juR/frame) 

45% 




{xl.4) 
111% 


{13.9 mg/cc) 
-30% 


{63 kVp) 
59% 


{173 vR/frame) 
81% 




{xl.6) 
186% 


{12.1 mg/cc) 
-40% 


{56kVp) 
55% 


{216 fiR/frame) 
135% 




{xl.8) 
276% 


{10.4 mg/cc) 


(52 kVp) 


{280 juR/frame) 




{x2.0) 


NOTE: Values in parenthesis indicate the threshold concentration and 
imaging techniques required to achieve the quoted change in threshold 
iodine contrast concentration. 





Table 5-11: Comparison of the Effects of Tube Voltage, Input Exposure and Geometric 
Magnification on the Surface Dose for a Range of Changes in Threshold Iodine 

Concentration at 440 jaR/frame 



Change in 

Threshold 

Concentration 


Change in SD with Change in SD with 
kVp II Exposure 


Change in SD with 
Magnification 


0.0% 




0.0% 


0.0% 




0.0% 


{10.2 mg/cc) 
10% 




{70kVp) 
21% 


{444 fjRJframe) 
47% 




{xl.2) 
45% 


{9.20 mg/cc) 
20% 




{65 kVp) 
49% 


{661 juR/frame) 
127% 




{xl.4) 
101% 


{8.18 mg/cc) 
30% 




{61 kVp) 
90% 


{1020 /uR/frame) 
271% 




{xl.6) 
168% 


{7.15 mg/cc) 
40% 




{56kVp) 
151% 


{1670 /jRj 'frame) 
554% 




{xl.8) 
247% 


{6.13 mg/cc) 




{50 kVp) 


{2940 /jR/frame) 




{x2.0) 


NOTE: 
imaging 


Values in parenthesis indicate the threshold concentration and 
; techniques required to achieve the quoted change in threshold 





109 



300% 



120 ^iR/frame 

a Magnification 
x Input Exposure 
• Tube Voltage 




-50% -40% -30% -20% -10% 
Change in Threshold Contrast 



0% 



600% 




-40% -30% -20% -10% 
Change in Threshold Contrast 



0% 



FIGURE 5-10: Change in surface dose versus change in threshold iodine concentration 
with tube voltage input exposure and magnification. 



110 

Energy Imparted 

The effects of tube voltage, input exposure and geometric magnification on 
energy imparted were demonstrated in Figure 5-4, Figure 5-6 and Figure 5-8, 
respectively. A comparison of the changes introduced to the energy imparted by varying 
these parameters to achieve a given change in threshold iodine concentration is given in 
Table 5-12 and Table 5-13. The changes listed in Table 5-12 have been computed using 
70 kVp, 119 pR/frame and xl.2 magnification as the starting point. The changes in 
Table 5-13 have been computed using 70 kVp, 444 pR/frame and xl.2 magnification as 
the starting point. 

Figure 5-1 1 shows the changes in energy imparted introduced by the tube voltage, 
input exposure and geometric magnification versus the corresponding changes in 
threshold iodine concentration. Unlike the case of surface dose, geometric magnification 
has a very limited effect on energy imparted and a linear effect on threshold contrast 
concentration (Figure 5-9). Thus, geometric magnification becomes the parameter of 
choice to improve signal detection with the least impact on energy imparted. The most 
inefficient parameter in terms of energy imparted to improve signal detection is the input 
exposure to the image intensifier. When quantum mottle is the primary noise source on 
an image, input exposure increases energy imparted two to three times more than tube 
voltage does for the same threshold contrast changes. As quantum mottle becomes less 
important to image noise, input exposure increases energy imparted dramatically without 
a significant increase in image quality. 



Ill 



Table 5-12: Comparison of the Effects of Tube Voltage, Input Exposure and Geometric 

Magnification on the Energy Imparted for a Range of Changes in Threshold Iodine 

Concentration at Low Input Exposures 



Change in 


Change in EI with 


Change in EI with 


Change in EI with 


Threshold 




kVp 




II Exposure 


Magnification 


Concentration 












0.0% 




0.0% 




0.0% 


0.0% 


(17.3 mg/cc) 




(70 kVp) 




(119 vR/frame) 


(xl.2) 


-10% 




9.6% 




19% 


3.9% 


(15.6 mg/cc) 




(67kVp) 




(142 jjR/frame) 


(xl.4) 


-20% 




21% 




45% 


7.7% 


(13.9 mg/cc) 




(63 kVp) 




(173 fjR/frame) 


(xl.6) 


-30% 




36% 




81% 


12% 


(12.1 mg/cc) 




(56 kVp) 




(216 jjR/frame) 


(xl.8) 


-40% 




56% 




135% 


15% 


(10.4 mg/cc) 




(52 kVp) 




(280 /uR/frame) 


(x2.0) 


NOTE: 


Values 


in parenthesis 


indicate the threshold concentration and 



imaging techniques required to achieve the quoted change in threshold 
iodine contrast concentration. 



Table 5-13: Comparison of the Effects of Tube Voltage, Input Exposure and Geometric 

Magnification on the Energy Imparted for a Range of Changes in Threshold Iodine 

Concentration at High Input Exposures 



Change in 


Change in EI with 


Change in EI with 


Change in EI with 


Threshold 


kVp 


II Exposure 


Magnification 


Concentration 








0.0% 


0.0% 


0.0% 


0.0% 


(10.2 mg/cc) 


(70kVp) 


(444 fjR/frame) 


(xl.2) 


10% 


15% 


47% 


1.6% 


(9.20 mg/cc) 


(65 kVp) 


(661 jjR/frame) 


(xl.4) 


20% 


33% 


127% 


3.1% 


(8.18 mg/cc) 


(61 kVp) 


(1020 fjR/frame) 


(xl.6) 


30% 


58% 


271% 


4.7% 


(7.15 mg/cc) 


(56 kVp) 


(1670 fjR/ frame) 


(xl.8) 


40% 


93% 


554% 


6.3% 


(6.13 mg/cc) 


(50kVp) 


(2940 jjR/frame) 


(x2.0) 


NOTE: 


Values in parenthesis indicate the threshold concentration and 



imaging techniques required to achieve the quoted change in threshold 
iodine contrast concentration. 



112 



200% 



120mR/frame 

x Input Exposure 
• Tube Voltage 
a Magnification 




-50% -40% -30% -20% -10% 
Change in Threshold Contrast 



0% 



600%, 



440 mR/frame 

x Input Exposure 
• Tube Voltage 
a Magnification 




70 

-50% 



-40% -30% -20% -10% 
Change in Threshold Contrast 



0% 



Figure 5-11: Change in energy imparted versus change in threshold iodine 
concentration with tube voltage input exposure and magnification. 



113 

Image Quality 

Figure 5-5 showed that tube voltage has a significant effect on the threshold 
iodine concentration. In both 120 (aR/frame and 440 \iR/ 'frame image intensifier input 
exposures, the tube voltage demonstrated a good power relationship of kVp" with 
threshold concentration, where n = 1.8. The high absorption of iodine at low tube 
voltages due to its 33 keV K-edge is the primary factor of decreased threshold contrast at 
low tube voltages. As the tube voltage decreases, the mean energy of the x-ray beam 
decreases from about 54 keV at 100 kVp (HVL = 5.34 mm Al) to 35 keV at 50 kVp 
(HVL = 2.65 mm Al). Thus, a 50 kVp x-ray beam with a mean energy just above the K- 
edge of iodine is expected to produce the highest image contrast. A second factor 
affecting contrast is scatter radiation. Scatter radiation increases with increased tube 
voltage and degrades contrast and the signal-to-noise ratio of the images. 

In Figure 5-7 where the threshold contrast is given as a function of input exposure 
to the image intensifier, two regions can be distinguished. The first region extends to an 
input exposure of about 250 ^R/frame and can be characterized as quantum noise limited 
(D'° 5 ), where quantum mottle is the primary source of image noise. Any increase to input 
exposure in this region will improve image quality by decreasing quantum mottle. 
Beyond the input exposure of 250 (iR/frame, the relationship between threshold iodine 
concentration and input exposure deviates from being quantum limited to D' 021 . At this 
region of input exposures, other noise factors (i.e., electronic noise, time jitter and 
structure noise) become equally important to quantum mottle. Such factors are not 
affected by any increase to input exposure. Additional increases to input exposure 
beyond 250 ^R/frame is not as effective in improving image quality. Thus, it is sensible 



114 

to select an input exposure of about 250 pR/frame for all image acquisitions in digital 
subtraction angiography. 

Figure 5-9 shows that geometric object magnification has a linear effect on 
threshold concentration. In both 120 pR/frame and 440 pR/frame input exposures, 
doubling geometric magnification decreased threshold concentration by a factor of two. 
The major factor that improves contrast with geometric magnification is the increase in 
the projected area of the imaged object (i.e., iodinated vessel). The increase of signal area 
improves the signal-to-noise ratio of the imaged object and improves signal detection. 
Another factor that improves threshold iodine concentration with increased magnification 
is the increased air-gap. Air-gaps act as scatter removal media. As the air-gap between 
the phantom and the image receptor increases with magnification, less scatter radiation 
reaches the image receptor and contrast is improved. 

The wide variability of readings at low input exposures may be explained by the 
significant image noise on those images, which makes image quality evaluation difficult. 
Possible variation of distance between the image and the observer will also introduce 
some variability during image evaluation. As the distance of the observer from the image 
varies, the perceived diameter of the vessel changes affecting signal detection (Burgess 
and Humphrey, 1997). 

Conclusions 
Reducing the x-ray tube voltage offered the largest improvement in image quality 
for a given increase in patient dose. Increasing the image intensifier input exposure 
beyond 250 pR/frame provided very little improvement in image quality. This image 



115 

intensifier input exposure level should not be exceeded in interventional neuroradiologic 
imaging, unless a valid justification is explicitly given. A linear relationship was 
observed between magnification and threshold concentration, which offers significant 
patient benefits when surface doses are not expected to exceed the threshold doses for the 
induction of deterministic effects. In cases when stochastic effects may be significant 
(i.e., pediatric cases), and deterministic effects can be tolerated, a significant increase in 
image quality can be achieved by increasing geometric magnification without introducing 
any significant changes to energy imparted. 

During digital subtraction angiography, the amount of iodine contrast that can be 
tolerated by a patient may also be of concern. Due to continuous dilution of the iodine 
contrast in the blood stream, the iodine concentration in the imaged blood vessel 
decreases significantly with distance. A higher iodine concentration would be required to 
image distant vessels especially during the capillary phase. Instead of increasing the 
iodine concentration in such cases, use of lower tube voltages will decrease the threshold 
concentration significantly to allow use of less of an amount of iodine. Changing the 
tube voltage from 70 kVp to 50 kVp, for example, will decrease the threshold 
concentration by more than 50%, which translates to 50% less iodine used on the patient. 
Such a decrease in tube voltage, however, results in about 140% increase in surface dose 
and 90% increase in energy imparted. 



CHAPTER 6 
CONCLUSIONS 



Patient Dosimetry 



Surface Doses 

Patients undergoing interventional neuroradiologic procedures receive significant 
radiation doses due to the complexity of such procedures and the amount of diagnostic 
information required to evaluate neuroradiologic abnormalities. The median values for 
the recorded surface doses to the 175 patients undergoing interventional neuroradiologic 
procedures were 1 .2 Gy in the frontal imaging plane (occipital region of the head) and 
0.64 Gy in the lateral imaging plane (next to temporal bone). The maximum surface dose 
received by a patient was of the order of 5 Gy in either imaging plane. The majority of 
the patients received surface doses between 0.3 Gy and 2.3 Gy. Thirty-three percent of 
the patients exceeded the dose taken to be the threshold surface dose (2 Gy) for 
deterministic injuries of the skin. Most of the surface dose was contributed by 
radiographic acquisitions, which accounted for the 67% of the total surface dose in the 
frontal plane and 78% of the total dose in the lateral plane. Fluoroscopy contributed only 
33% and 22% of the total surface dose in the frontal and lateral planes, respectively. 

Although about 33% of the patients in this study may have exceeded the nominal 
threshold surface dose to the skin for the induction of deterministic effects, there were no 

116 



117 

cases of epilation or skin erythema observed among the 1 75 patients who underwent 
interventional neuroradiologic procedures. When determining the likelihood of 
deterministic effects to patients undergoing neuroradiologic examinations, it should be 
noted that the radiation doses are delivered over an extended time period, which may be 
as long as several hours and the radiation field may vary over the patient. For acute 
radiation exposures, observable effects such as total epilation are more likely to occur at 
doses much higher than 2 Gy, which is considered to be the threshold for the induction of 
deterministic effects. 



Effective Doses 

Patients who underwent interventional neuroradiologic procedures received 
typically between 10 mSv and 70 mSv of effective dose. The distribution of effective 
doses to these patients had a median value of 36 mSv with a maximum effective dose of 
156 mSv. Patient who underwent diagnostic procedures had a median value of 38 mSv 
total effective dose. The median value of the total effective dose to patients who 
underwent therapeutic embolization procedures was 74 mSv. Ten of the seventeen adult 
embolization procedures exceeded the 70 mSv value of total effective dose. Two thirds 
of the effective dose was contributed by imaging in the frontal imaging plane. As was the 
case for surface dose, radiographic acquisitions accounted for the largest fraction of the 
effective dose to adult patients. On average, about 64% of the effective dose 
corresponded to radiographic acquisitions. Use of fluoroscopy accounted for only one 
third of the total effective dose received by patients during interventional neuroradiologic 
procedures. 



118 

Using the recent risk coefficient of 5x1 0" 5 cancers per mSv of effective dose (ICRP 
1991), the effective doses computed in this study can be converted to a stochastic 
detriment. The median effective dose of 36 mSv to adult patients undergoing 
interventional neuroradiologic procedures would result in one fatal cancer for every 555 
such procedures. An effective dose of 74 mSv to adults undergoing therapeutic 
embolization procedures would result in one fatal cancer for every 270 embolizations. In 
general, these stochastic risks are low compared to the life saving benefits the patients 
receive by undergoing interventional neuroradiologic procedures. 

In the case of pediatric patients undergoing interventional neuroradiologic 
procedures, the effective dose was found to have a linear correlation to patient mass 
(age). This correlation resulted to higher effective doses to pediatric patients compared to 
the corresponding adult effective doses. The median effective dose to pediatric patients 
was 44 mSv. Pediatric patients undergoing embolization procedures received a median 
effective dose of 66 mSv. 

Any resultant stochastic detriment depends on the age of the exposed individual. 
The stochastic radiation risks of carcinogenesis and genetic effects are generally greater 
for children than for adults to at least a factor of two (ICPR 1991, NCRP 1985). These 
factors would need to be taken into account when converting any pediatric effective doses 
into a value of risk or detriment. In general, pediatric patients receive approximately the 
same effective doses as adult patients do from interventional neuroradiologic procedures. 
The fact that stochastic risks associated with children are higher than the risks associated 
with adults requires that we should focus our attention to reducing doses to pediatric 
patients undergoing interventional neuroradiologic procedures. 



119 

Image Quality 

The tube voltage had the strongest effect on image quality for the same increase of 
patient dose among the three technique parameters studied — tube voltage, image 
intensifier input exposure and geometric object magnification. The photoelectric effect of 
iodine at low tube voltages improves contrast detectability and thus signal detection as 
tube voltage decreases. This strong correlation between tube voltage and threshold 
concentration can be expressed as a power relationship of kVp" where n « 1.8. Tube 
voltage also has a significant effect on patient doses. A decrease of 70%-80% in surface 
dose and effective dose can be achieved by increasing the tube voltage from 50 kVp to 
100 kVp while the other two parameters remain constant, with the largest decrease in 
doses occurring between 50 kVp and 66 kVp. 

Two regions of image intensifier input exposures were observed regarding changes 
in threshold concentration. The first region was extended below 250 p.R/frame and was 
quantum mottle limited. Increasing input exposure by a factor of two in this region 
improved threshold concentration by 34%. The second region was observed above 250 
(jR/frame, where less improvement in image quality could be achieved beyond this point, 
since quantum mottle did not appear to be the primary source of image noise. An 
improvement of only 17% could be achieved by doubling the input exposure at the region 
beyond 250 p.R/frame. Thus, it is sensible to select an input exposure of 250 p.R/frame as 
the input exposure for all image acquisitions in digital subtraction angiography to 
capitalize in the improvement of image quality without significantly increasing patient 
doses. 



120 

Geometric object magnification was found to be linearly correlated to threshold 
contrast. Increasing geometric magnification by a factor of two improved image quality 
by decreasing threshold iodine contrast by a factor of two. This increase in image quality 
by applying geometric magnification can be achieved with a minimal increase in effective 
dose to the patient. Thus, magnification can be used to improve signal detection where 
stochastic risks may be of concern. On the other hand, geometric magnification had a 
strong effect on surface dose. Geometric magnification should be minimized where there 
is a concern of deterministic injuries. 

In general, the three imaging parameters studied in this work to quantify their 
effects on image quality and patient dose indicated that reducing the x-ray tube voltage 
offered the largest improvement in image quality for a given increase in patient dose. 
Increasing the image intensifier input exposure beyond 250 uR/frame provided very little 
improvement in image quality. A linear relationship was observed between 
magnification and threshold concentration, which offers significant patient benefits when 
surface doses are not expected to exceed the threshold doses for the induction of 
deterministic injuries. These facts should be considered every time a selection of imaging 
techniques is required for optimization purposes. 



Future Work 

The traditional threshold for the induction of deterministic injuries to the skin was 

proposed to be of the order of 2 Gy (Wagner et ah, 1994). A significant number (28%) of 

patients undergoing interventional neuroradiologic procedures studied in this work 

exceeded the proposed threshold for deterministic effects without noticing any radiation 



121 

injuries such as epilations and erythemas. This suggests that a higher threshold value 
should be considered as the triggering point of such deterministic injuries. More work is 
required to re-evaluate these thresholds and determine a better value. 

In this work, there was no image processing other than window and level applied to 
the digitally subtracted images during the evaluation of image quality. A multitude of 
processing algorithms and techniques are available that may be able to improve image 
quality without affecting patient dose. Such image processing techniques may be studied 
to evaluate their effects on signal detection as they may apply on digital subtraction 
angiography. 

Digital subtraction angiography uses the photoelectric effect in iodine occurring at 
33 keV to differentiate between angiographic structures and other anatomy. Different 
contrast agents with higher atomic numbers and K-edge energies will be able to reduce 
patient dose for the same image quality. However, the toxicity of higher atomic number 
agents may limit such effects. Such contrast agents need to be evaluated for applications 
in interventional neuroradiology to quantify their effects on image quality and the 
corresponding changes in patient dose. 

The already high quantum detective efficiency (DQE) of current image intensifier 
systems used in x-ray imaging sets a limit to further improvement in signal detection with 
the current imaging equipment used in interventional neuroradiology. It is likely that a 
direct digital detector that will be able to exceed the efficiency of current image 
intensifiers, improve signal detection and decrease patient dose will replace the image 
intensifier in the future. The application of such digital x-ray image receptors requires 
further development and evaluation. 



BIBLIOGRAPHY 



Almen A, Mattsson S. "On the calculation of effective doses to children and 
adolescents." Journal of Radiation. Protection. 16: 81-89 (1996). 

Arnold BA, Scheibe PO. "Noise analysis of a digital radiography system." American 
Journal of Roentgenology 142: 609-613 (1984). 

Baiter S, Ergun D, Tscholl E, Buchmann F, Verhoeven L. "Digital subtraction 
angiography: fundamental noise characteristics." Radiology 152: 195-198 (1984). 

Baxter WT, Davidenko JM, Loew LM, Wuskell JP, Jalife J. "Technical features of a 
CCD video camera system to record cardiac fluorescence data." Ann Biomed Eng 
25: 713-725 (1997). 

Bengtsson G, Blomgren P, Bergman K, Aberg L. "Patient exposures and radiation risks 
in Swedish diagnostic radiology." Acta Radiologica Oncology 17: 81-105 (1978). 

Bergeron P, Carrier R, Roy D, Blais N, Raymond J. "Radiation doses to patients in 
neurointerventional procedures." American Journal of Neuroradiology 15: 1809- 
1812(1994). 

Berthelsen B, Cederblad A. "Radiation doses to patients and personnel involved in 
embolization of intracerebral arteriovenous malformations." Acta Radiologica 
32: 492-497 (1991). 

Boone JM. "Parametrized x-ray absorption in diagnostic radiology from Monte Carlo 
calculations: Implications for x-ray detector design." Medical Physics 19: 1467- 
1473 (1992). 

Boone JM, Shaber GS, Tecotzky M. "Dual-energy mammography: a detector analysis." 
Medical Physics 17: 665-675 (1990). 

Burgess AE. "Effect of quantization noise on visual signal detection in noisy images." J 
Opt SocAm [A] 2: 1424-1428 (1985). 

Burgess AE, Humphrey K. "Visual perception limits in angiography." Proceedings of 
the Society of Photo-Optical instrumentation Engineers 127: 51-59 (1997). 

Burkhart RL. Patient Radiation Exposures in Diagnostic Radiology Examinations: An 
Overview. HHS Publication (FDA), Rockville, MD 82-8219(1983). 

122 



123 



Cameron, J. "Dose equivalent out — Imparted energy in." HPS Newsletter 7-8 (1992). 

Carlsson C. "Determination of integral absorbed dose from exposure measurements." 
Acta Radiologica 1: 433-458 (1963). 

Carlsson C. "Integral absorbed doses in roentgen diagnostic procedures. I. The 
dosimeter." Acta Radiologica Therapy Physics Biology 3: 310-326 (1965a). 

Carlsson, C. "Integral absorbed doses in roentgen diagnostic procedures. II. 
Measurement of integral doses in two roentgen diagnostic departments." Acta 
Radiologica Therapy Physics Biology 3: 384-408 (1965b). 

Carlsson G, Carlsson C. "Relations between effective dose equivalent and mean 
absorbed dose (energy imparted) to patients in diagnostic radiology." Physics in 
Medicine and Biology 3 1 : 9 1 1 -92 1 ( 1 986). 

Carlsson G, Carlsson C, Persliden, J. "Energy imparted to the patients in diagnostic 
radiology: calculation of conversion factors for determining the energy imparted 
from measurements of the air collision kerma integrated over beam area." Physics 
in Medicine and Biology 29: 1329-1341 (1984). 

Chappel CL, Broadhead DA, Faulkner K. "A phantom based method for deriving typical 
patient doses from measurements of dose-area product on populations of 
patients." British Journal of Radiology 68: 1083-1086 (1995). 

Chappie C, Faulkner K, Hunter E. "Energy imparted to neonates during X-ray 
examinations in a special care baby unit." British Journal of Radiology 67: 366- 
370(1994). 

Chopp M, Portnoy HD, Schuring R, Croisaant P. "Clinical dosimetry during cerebral 
arteriography." Neuroradiology 20: 79-81 (1980). 

Cohen G, Wagner L, Rauschkolb E. "Evaluation of a digital subtraction angiography 
unit." Radiology 144: 613-617 (1982). 

Doi K, Imhof K. "Noise reduction by radiographic magnification." Radiology 122: 479- 
487(1977). 

Doi K. Rossmann K. "The effect of radiographic magnification on blood vessel imaging 
with various screen-film systems." Medical Physics 1: 257-261 (1974). 

Esthappan J, Harauchi H, Hoffmann KR. "Evaluation of imaging geometries calculated 
from biplane images." Medical Physics 25: 965-975 (1998). 

Faulkner K, Harrison R. "Estimation of effective dose equivalent in diagnostic 
radiology." Physics in Medicine and Biology 33: 83-91 (1988). 



124 



Feygelman VM, Huda W, Peters KR. "Effective dose equivalents to patients undergoing 
cerebral angiography." American Journal of Neuroradiology 13: 845-849 (1992). 

Field SB, Upton AC. "Non-stochastic effects: Compatibility with present ICRP 
recommendations." Int J Radial Biol 48: 81-94(1985). 

Fry RJM. "Effects of low doses of radiation." Health Physics 70: 823-827 (1996). 

Fujita H, Doi K, Giger ML. "Investigation of basic imaging properties in digital 
radiography. 6. MTFs of II-TV digital imaging systems." Medical Physics 12: 
713-720(1985). 

Gibbs S, Pujol A, Chen T, Malcolm A, James E. "Patient risk from interpatient 
radiography." Oral Surgery 58: 347-354(1984). 

Giger ML. Investigation of Basic Imaging Properties in Digital Radiography. Ph.D. 
Dissertation, University of Chicago, Chicago, IL (1985). 

Giger ML, Doi K. "Investigation of basic imaging properties in digital radiography. 1 . 
Modulation transfer function." Medical Physics 11: 287-295 (1984). 

Giger ML, Doi K, Fujita H. "Investigation of basic imaging properties in digital 
radiography. 7. Noise Wiener spectra of II-TV digital imaging systems. " 
Medical Physics 13: 131-138 (1986a). 

Giger ML, Ohara K, Doi K. "Investigation of basic imaging properties in digital 
radiography. 9. Effect of displayed gray levels on signal detection." Medical 
Physics 13: 312-318 (1986b). 

Gkanatsios NA. Computation of Energy Imparted in Diagnostic Radiology. Master 
Thesis University of Florida, Gainesville, FL (1995). 

Gkanatsios NA, Huda W. "Computation of energy imparted in diagnostic radiology." 
Medical Physics 24: 571-579 (1997). 

Gkanatsios NA, Huda W, Peters KR, Freeman JA. "Evaluation of an on-line patient 
exposure meter in neuroradiology." Radiology 203: 837-842 (1997). 

Grosswendt B. "Backscater factors for x-rays generated at voltages between 10 and 100 
kV." Physics in Medicine and Biology 29:579-591 (1984). 

Grosswendt B. "Dependence of the photon backscatter phantom for water on source-to- 
phantom distance and irradiation field size." Physics in Medicine and Biology 
35:1233-1245(1990). 

Harrison R. "A re-evaluation of the 'saturated scatter' method for estimating the energy 
imparted to patients during diagnostic radiology examinations." Physics in 
Medicine and Biology 28: 701-707 (1983). 



125 



Hart D, Jones DG, Wall BF. NRPB Report 262: Estimation of Effective Dose in 
Diagnostic Radiology from Entrance Surface Dose and Dose-Area Product 
Measurements. National Radiological Protection Board, Chilton, Oxon, UK 
(1994a). 

Hart D, Jones DG, Wall BF. NRPB Software SR262: Normalized Organ Doses for 
Medical X-Ray Examinations Calculated Using Monte Carlo Techniques. 
National Radiological Protection Board, Didcot, Oxon (1994b). 

Hart D, Jones DG, Wall BF. NRPB Software SR279: Normalized Organ Doses for 
Pediatric X-Ray Examinations Calculated Using Monte Carlo Techniques. 
National Radiological Protection Board, Didcot, Oxon (1996a). 

Hart D, Jones DG, Wall BF. NRPB Report R279: Coefficients for Estimating Effective 
Doses from Pediatric X-Ray Examinations. National Radiological Protection 
Board, Didcot, Oxon (1996b). 

Hasagawa BH. The Physics of Medical X-Ray Imaging. Medical Physics Publishing, 
Madison, WI (1991). 

Haus AG (editor). AAPM Monogram #2. The Physics of Medical Imaging: Recording 
System Measurements and Techniques. American Institute of Physics, New York, 
NY (1979). 

Huda W. "Is energy imparted a good measure of the radiation risk associated with CT 
examinations?" Physics in Medicine and Biology 29: 1 137-1 142 (1984). 

Huda W, Bews J. "Population irradiation factors (PIFs) in diagnostic medical 
dosimetry." Health Physics 59: 345-347 (1990). 

Huda W, Bissessur K. "Effective dose equivalents, H E , in diagnostic radiology." 
Medical Physics 17: 998-1003 (1990). 

Huda W, Gkanatsios NA. "Effective doses and energy imparted in diagnostic radiology." 
Medical Physics 24: 1311-1316 (1997). 

Huda W, Gkanatsios NA. "Effective doses associated with extremity radiographic 
examinations." Health Physics 75: 492-499 (1998). 

Huda W, Lentle B, Sutherland J. "The effective dose equivalent in radiology." Journal 
of the Canadian Association of Radiologists 40: 3-4 (1989a). 

Huda W, McLellan J, McLellan Y. "How will the new definition of 'effective dose' 
modify estimates of dose in diagnostic radiology?" Journal of Radiation 
Protection 11: 241-247 (1991). 

Huda W, Peters KR. "Radiation-induced temporary epilation after a neuroradiologically 
guided embolization procedure." Radiology 193:642-644 (1994). 



126 



Huda W, Sandison G, Palser R, Savoie D. "Radiation doses and detriment from chest x- 
ray examinations." Physics in Medicine and Biology 34: 1477-1492 (1989b). 

Hummel R, Wesenberg R, Amundson G. "A computerized x-ray dose-monitoring 
system." Radiology 156: 231-234 (1985). 

lies WJ. NRPB Report R204: The Computation of Bremsstrahlung X-Ray Spectra Over 
an Energy Range 15 keV to 300 keV. National Radiological Protection Board, 
Didcot, Oxon (1987). 

Institute of Physical Sciences in Medicine. Report 55: Are X-Rays Safe Enough? 
Patient Doses and Risks in Diagnostic Radiology. The Institute of Physical 
Sciences in Medicine, York, England (1988). 

International Commission on Radiological Protection (ICRP). Publication 23: Report of 
the Task Group on Reference Man. Pergamon Press, Oxford (1975). 

International Commission on Radiological Protection (ICRP). Publication 26: Annals of 
the ICRP 1(3). Recommendations of the International Commission on 
Radiological Protection. Pergamon Press, Oxford, England (1977). 

International Commission on Radiological Protection (ICRP). Annals of the ICRP 2(1). 
Statement for the 1978 Stockholm Meeting of the International Commission on 
Radiological Protection. Pergamon Press, Oxford, England (1978). 

International Commission on Radiological Protection (ICRP). Publication 34: Protection 
of the Patient in Diagnostic Radiology. Pergamon Press, Oxford (1982). 

International Commission on Radiological Protection (ICRP). Publication 36: Annals of 
the ICRP 10(2/3). Cost-Benefit Analysis in the Optimization of Radiation 
Protection. Pergamon Press, Oxford (1983). 

International Commission on Radiological Protection (ICRP). Publication 47: Annals of 
the ICRP 14(3). Non-Stochastic Effects of Ionizing Radiation. Pergamon 
Press,Oxford(1984). 

International Commission on Radiological Protection. Publication 53: Radiation Doses 
to Patients from Radiopharmaceuticals, Annals of the ICRP 18 No 1-4. 
Pergamon Press, Oxford, England (1987). 

International Commission on Radiological Protection (ICRP). Publication 60: Annals of 
the ICRP 21(1-3). 1990 Recommendations of the International Commission on 
Radiological Protection. Pergamon Press, Oxford (1991). 

Johns HE, Cunningham JR. The Physics of Radiology, 4th Edition. Charles C Thomas, 
Springfield, IL (1983). 



127 



Jones D, Wall B. NRPB Report 186: Organ Doses from Medical X-Ray Examinations 
Calculated Using Monte Carlo Techniques. National Radiological Protection 
Board, Chilton, UK (1985). 

Klevenhagen SC. "The built-up of backscatter at the energy range 1 mm Al to 8 mm Al 
HVL." Physics in Medicine and Biology 27: 1035-1043(1982). 

Koedooder K, Venema HW. "Filter materials for dose reduction in screen-film 
radiography." Physics in Medicine and Biology 31: 585-600 (1986). 

Le Heron J. "Estimation of effective dose to the patient during medical x-ray 
examinations from measurements of the dose-area products." Physics in 
Medicine and Biology 37: 21 17-2126 (1992). 

Lindskoug BA. "The reference man in diagnostic radiology dosimetry." British Journal 
of Radiology 65: 431-437 (1992). 

Manny EF, Burkhart RL. Measurement Techniques for Use With Technique/Exposure 
Guides. HHS Publication (FDA) 85-8248, Rockville, MD (1985). 

McParland BJ. "A study of patient radiation doses in interventional radiological 
procedures." British Journal of Radiology 71: 175-185(1998). 

Merriam GR Jr, Focht EF. "A clinical study of radiation cataracts in relationship to 
dose." American Journal of Roentgenology 11: 759-785 (1957). 

Metz CE, Doi K. "Transfer function analysis of radiographic imaging systems." Physics 
in Medicine and Biology 24: 1079-1 106 (1979). 

Nagel HD. "Comparison of performance characteristics of conventional and K-edge 
filters in general diagnostic radiology." Physics in Medicine and Biology 34: 
1269-1287 (1989). 

National Academy of Sciences. BEIR V: Health Effects of Exposure to Low Levels of 
Ionizing Radiation. National Academy Press, Washington DC (1990). 

National Council on Radiation Protection and Measurements. Report No. 80: Induction 
of Thyroid Cancer by Ionizing Radiation. National Council on Radiation 
Protection and Measurements, Bethesda, MD (1985). 

National Council on Radiation Protection and Measurements. Report No. 100: Exposure 
of the US Population from Diagnostic Medical Radiation. National Council on 
Radiation Protection and Measurements, Bethesda, MD (1989). 

Norbash AM, Busick D, Marks MP. "Techniques for reducing interventional 
neuroradiologic skin dose: tube position rotation and supplemental beam 
filtration." American Journal of Neuroradiology 17: 41-49 (1996). 



128 



Nuclear Regulatory Commission. 10CFR19: Notices, Instructions, and Reports to 
Workers: Inspection and Investigations. Nuclear Regulatory Commission, 
Washington, DC (1995 a). 

Nuclear Regulatory Commission. 10CFR20: Standards for Protection Against 
Radiation. Nuclear Regulatory Commission, Washington, DC (1995b). 

Padovani R, Contento G, Fabretto M, Malisan MR, Barbina V, Gozzi G. "Patient doses 
and risks from diagnostic radiology in North-east Italy." British Journal of 
Radiology 60: 155-165 (1987). 

Patrocinio HJ, Bissonnette JP, Bussiere MR, Schreiner LJ. "Limiting values of 
backscatter factors for low-energy x-ray beams." Physics in Medicine and 
Biology 41: 239-253 (1996). 

Perslider J, Carlsson G. "Energy imparted to water slabs by photons in energy range 5- 
300 keV. Calculations using Monte Carlo photon transport model." Physics in 
Medicine and Biology 29: 1075-1088 (1984). 

Puskin JS, Nelson CB. "Estimates of radiogenic cancer risks." Health Physics 69: 93- 
101 (1996). 

Rajapakshe R, Shalev S. "Noise analysis in real-time portal imaging. I. Quantization 
noise." Medical Physics 21: 1263-1268 (1994). 

Roehrig H, Nudelman S, Fisher HD, Frost MM, Capp MP. "Photoelectronic imaging for 
radiology." IEEE Transactions on Nuclear Science 28: 190-204 (1981). 

Rubin P, Casarett GW. Clinical Radiation Pathology. Vol I and II. WB Saunders 
Philadelphia, PA (1968). 

Rudin S, Bednarek DR. "Non-hardening x-ray beam attenuating filters." Abstract, 
AAPM Annual Meeting Program. Medical Physics 20:918 (1993). 

Sandor T, Natt P. "Effect of radiographic magnification on image contrast of blood 
vessels." American Journal of Roentgenology 134: 159-192(1980). 

Sandrik JM, Wagner RF. "Absolute measures of physical image quality: measurement 
and application to radiographic magnification." Medical Physics 9: 540-549 
(1982). 

Shleien B (editor). The Health Physics and Radiological Health Handbook. Scinta, Inc, 
Silver Spring, MD (1992). 

Shope TB. "Radiation-induced skin injuries from fluoroscopy." Radiographics 16: 
1195-1199(1996). 



129 



Shrimpton P, Wall B. "An evaluation of the Diamentor transmission ionization chamber 
in indicating exposure-area product (R cm 2 ) during diagnostic radiological 
examinations." Physics in Medicine and Biology 27: 871-878 (1982). 

Shrimpton PC, Jones DJ, Wall BF. "The influence of tube filtration and potential on 
patient dose during x-ray examinations." Physics in Medicine and Biology 33: 
1205-1212(1988). 

Shrimpton P, Wall B, Jones D, Fisher E. "The measurement of energy imparted to 
patients during diagnostic x-ray examinations using the Diamentor exposure-area 
product meter." Physics in Medicine and Biology 19: 1 199-1208 (1984). 

Suleiman OH, Antonsen R, Conway B, McGrohan J, Rueter F, Slayton R. "Assessing 
patient exposure in fluoroscopy." Radiation Protection and Dosimetry 43:251- 
252 (1992). 

Tapiovaara MJ, Sandborg M. "Evaluation of image quality in fluoroscopy by 
measurements and Monte Carlo calculations." Physics in Medicine and Biology 
40: 589-607 (1995). 

Thompson WM, Halvorsen RA, Gedgaudas RK, Kelvin FM, Rice RP, Woodfield S, 
Johnson GA, Hedlung LW, Jorgensen DB. "High kVp vs. low kVp for T-tube 
operative cholangiography." Radiology 146: 635-642 (1983). 

Trout ED, Kelley JP, Cathey GA. "The use of filters to control radiation exposure to the 
patient in diagnostic roentgenology." American Journal of Roentgenology, 
Radium Therapy and Nuclear Medicine 57: 946-963 (1952). 

Tucker M, Barnes G, and Chakraborty D. "Semi-empirical model for generating tungsten 
target x-ray spectra." Medical Physics 18: 211-218 (1991). 

United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR). 
Report to the General Assembly: Sources, Effects and Risks of Ionizing Radiation. 
United Nations, New York, NY (1988). 

United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR). 
1993 Report to the General Assembly: Medical Radiation Exposures. United 
Nations, New York, NY (1993). 

Vano E, Arranz L, Sastre JM, Moro C, Ledo A, Garate MT, Minguez I. "Dosimetric and 
radiation protection considerations based on some cases of patient skin injuries in 
interventional cardiology." British Journal of Radiology 71: 510-516 (1998). 

Wagner, L. "Absorbed dose in imaging: Why measure it?" Radiology 178: 622-623 
(1991). 



130 



Wagner LK, Cohen G, Wong WH, Amtey SR. "Dose efficiency and the effects of 
resolution and noise on detain perceptibility in radiographic mammography." 
Medical Physics 8: 24-32 (1981a). 

Wagner LK, Cohen G, Wong WH, Amtey SR. "Resolution in radiographic 
magnification." Medical Physics 8: 382-384 (1981b). 

Wagner LK, Eifel PJ, Geise RA. "Potential biological effects following high x-ray dose 
interventional procedures." JofVasclntRad 5:71-84(1994). 

Wall B, Fisher E, Paynter R, Hudson A, Bird P. "Doses to patients from 
phantomographic and conventional dental radiography." British Journal of 
Radiology 52: 727-734 (1979). 

Wall B, Harrison R, Spiers F. ISPM Report #53: Patient Dosimetry Techniques in 
Diagnostic Radiology. The Institute of Physical Science in Medicine, York, 
England (1988). 



BIOGRAPHICAL SCHETCH 

Nikolaos A. Gkanatsios was born in Greece in 1970 to Argyrios and Anastasia 
Gkanatsios. He came to the United States in 1988 and earned a bachelor's degree in 
nuclear engineering from Worcester Polytechnic Institute, Worcester, MA, in 1992. The 
same year he joined the graduate program of the University of Florida, where he earned 
his Master of Science in medical physics in August of 1995. He was accepted as a Ph.D 
candidate in the Department of Nuclear and Radiological Sciences at the University of 
Florida. Nikolaos Gkanatsios earned his Doctor of Philosophy degree from the 
University of Florida in 1998. 






131 



I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of PI 




JUiyL/ 



2S S. Tulenko, Chair 
lessor of Nuclear and Radiological 
Engineering 



I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. / [ 





v^ 



Walter Huda, Cochair 
Associate Professor of Nuclear and 
Radiological Engineering 



I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. 




Wesley E. I^blch 
Associate Professor of Nuclear and 
Radiological Engineering 

I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. 





Janice C. Honeyman 

Associate Professor of Computer and 
Information Science and Engineering 

I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. 




^KeithRPetefT 
Associate Professor of Radiology 



I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. 

Irvin F. Hawkins 
Professor of Radiology 

This dissertation was submitted to the Graduate Faculty of the College of 
Engineering and to the Graduate School and was accepted as partial fulfillment of the 
requirements for the degree of Doctor of Philosophy^ 

December, 1998 

Winfred M. Phillips 

Dean, College of Engineering 



M.J. Ohanian 

Dean, Graduate School 







LD 
1780 

* c n53f} 









UNIVERSITY OF FLORIDA 



3 1262 08555 1009